* server: ci: windows build and tests
* server: ci: remove tmp push branch
* server: ci: EOF EOL
* Use builti
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* server: tests: server graceful shutdown, then kill, then hard kill
* server: tests: remove python2 unicode string
* server: tests: remove wrong comment on server starting, close_fds is always true
* server: tests: server kill, if pid exists
* server: tests: remove dependency to killall
* server: tests: ci windows: pid exists better handling
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* add gritlm example
* gritlm results match
* tabs to spaces
* comment out debug printing
* rebase to new embed
* gritlm embeddings are back babeee
* add to gitignore
* allow to toggle embedding mode
* Clean-up GritLM sample code.
* Fix types.
* Flush stdout and output ending newline if streaming.
* mostly style fixes; correct KQ_mask comment
* add causal_attn flag to llama_cparams
* gritml : minor
* llama : minor
---------
Co-authored-by: Douglas Hanley <thesecretaryofwar@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: bench: Init a bench scenario with K6
See #5827
* server: bench: EOL EOF
* server: bench: PR feedback and improved k6 script configuration
* server: bench: remove llamacpp_completions_tokens_seconds as it include prompt processing time and it's misleading
server: bench: add max_tokens from SERVER_BENCH_MAX_TOKENS
server: bench: increase truncated rate to 80% before failing
* server: bench: fix doc
* server: bench: change gauge custom metrics to trend
* server: bench: change gauge custom metrics to trend
server: bench: add trend custom metrics for total tokens per second average
* server: bench: doc add an option to debug http request
* server: bench: filter dataset too short and too long sequences
* server: bench: allow to filter out conversation in the dataset based on env variable
* server: bench: fix assistant message sent instead of user message
* server: bench: fix assistant message sent instead of user message
* server : add defrag thold parameter
* server: bench: select prompts based on the current iteration id not randomly to make the bench more reproducible
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* perplexity : support using multiple sequences to allow larger batch sizes
ggml-ci
* set cparams.n_parallel to the number of sequences
* print tested n_ctx, add assert
* add cmake build toggle to enable ssl support in server
Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
* add flags for ssl key/cert files and use SSLServer if set
All SSL setup is hidden behind CPPHTTPLIB_OPENSSL_SUPPORT in the same
way that the base httlib hides the SSL support
Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
* Update readme for SSL support in server
Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
* Add LLAMA_SERVER_SSL variable setup to top-level Makefile
Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
---------
Signed-off-by: Gabe Goodhart <ghart@us.ibm.com>
* mamba : begin working on support for Mamba SSM
* mamba : begin figuring out how to (ab)use the kv cache for Mamba
* mamba : recurrent inference almost works, but incoherent
* mamba : recurrent inference WORKS!!!
* convert : optionally use d_conv and d_state from config.json for Mamba
* mamba : refactor recurrent conv, resulting in 20% perf increase
It's still slower than I'd like, but I did not really optimize `ggml_exp` yet.
I also refactored `ggml_exp` to work with tensors with more than 2 dimensions.
* ggml : parallelize ggml_exp
This results in 8% faster token generation for Mamba-130M.
* mamba : simplify the conv step with a self-overlapping view
Turns out the conv_state can be made smaller by one column.
Note that this breaks existing GGUFs of Mamba,
because the key_value_length field is tied to the conv_state size.
Convolution with a self-overlapping view is cool!
And it's much simpler than what I initially thought would be necessary
to make the convolution step work with more than 1 token at a time.
Next step is to make the SSM step work on batches of tokens too,
and thus I need to figure out a way to make a parallel selective scan
which will keep the ssm_state small and won't make it bigger
by a factor of (n_layer * batch_size).
* llama : fix Mamba KV self size wrongly displaying as f16 instead of f32
Relatedly, I also tried to see if other types than f32 worked for the states,
but they don't, because of the operators used.
It's probably better anyway to keep lots of precision there,
since the states are small anyway.
* mamba : fix self-overlapping view depth stride
* mamba : handle batches of more than 1 token
This means running Mamba no longer crashes when using the default settings!
And probably also slightly faster prompt processing.
Both batched and non-batched processing yield the same output.
Previously, the state was not cleared when starting a sequence.
Next step is to make the KV cache API work as expected for Mamba models.
* ggml: add ggml_ssm_scan to help with parallel selective scan
If the selective scan was implemented without a custom operator,
there would be waaay too many nodes in the graph. For example,
for Mamba-130M, with a batch size of 512 (the default),
a naive selective scan could add at least 24*512=12288 nodes,
which is more than LLAMA_MAX_NODES (8192),
and that's only for the smallest Mamba model.
So it's much cleaner with a custom operator.
Not sure about the name, though.
* ggml : in ggml_ssm_scan, merge multiple rows in the same vec operation
This will help with performance on CPU if ggml_vec_mul_f32
and ggml_vec_add_f32 are ever optimized with SIMD.
* mamba : very basic quantization support
Mostly works, but there is currently no difference
between the variants of a k-quant (e.g. Q4_K_S and Q4_K_M are the same).
Most of the SSM-specific weights can be kept in f32 without affecting
the size that much, since they are relatively small.
(the linear projection weights are responsible for most of Mamba's size)
Too much quantization seems to make the state degrade quite fast, and
the model begins to output gibberish.
It seems to affect bigger models to a lesser extent than small models,
but I'm not sure by how much.
Experimentation will be needed to figure out which weights are more important
for the _M (and _L?) variants of k-quants for Mamba.
* convert : fix wrong name for layer norm weight of offical Mamba models
I was using Q-bert/Mamba-* models before, which have a slighlty different
naming scheme for the weights.
(they start with "model.layers" instead of "backbone.layers")
* mamba : fuse more steps of the SSM scan in the ggml_ssm_scan operator
This increases performance on CPU by around 30% for prompt processing,
and by around 20% for text generation.
However, it also makes the ggml_exp and ggml_soft_plus operators unused.
Whether or not they should be kept will be decided later.
* convert : for Mamba, also consider the "MambaLMHeadModel" arch name
It's the name of the class of the official implementation,
though they don't use it (yet) in the "architectures" field of config.json
* mamba : fix vocab size problems with official models
The perplexity was waaaay to high for models with a non-round vocab size.
Not sure why, but it needed to be fixed in the metadata.
Note that this breaks existing GGUF-converted Mamba models,
but **only if** the vocab size was not already rounded.
* ggml : remove ggml_exp and ggml_soft_plus
They did not exist anyway outside of this branch,
and since ggml_ssm_scan fused operations together, they are unused.
It's always possible to bring them back if needed.
* mamba : remove some useless comments
No code change.
* convert : fix flake8 linter errors
* mamba : apply suggestions from code review
* mamba : remove unecessary branch for row-wise ssm_state and C multiplication
It was previously done to avoid permuting when only one token is processed
at a time (like when generating text), but permuting is cheap,
and dynamically changing the compute graph is not future-proof.
* ggml : in ggml_ssm_scan, use more appropriate asserts
* ggml : rename the destination pointer in ggml_compute_forward_ssm_scan_f32
* mamba : multiple sequences, but one at a time
This is a step towards making this Mamba implementation usable
with the server example (the way the system prompt is kept when clearing
the client slots will need to be changed before this can work, though).
The KV cache size for this kind of model is tied to the maximum number
of sequences kept at any single time.
For now, this number is obtained from n_parallel (plus one,
to have an extra sequence to dedicate to the system prompt),
but there might be a better way to do this which won't also
make the main example use 2 cells even if only 1 is really used.
(for this specific case, --parallel 0 helps)
Simultaneous sequence processing will probably require changes to
ggml_ssm_scan, and possibly a new operator for the conv step.
* mamba : support llama_kv_cache_seq_cp
This (mis)uses the logic around K shifts, because tokens in a state
can't be shifted anyway, and because inp_K_shift has the right shape and type.
Using ggml_get_rows is a nice way to do copies, but copy chains can't work.
Fortunately, copy chains don't really seem to be used in the examples.
Each KV cell is dedicated to the sequence ID corresponding to its own index.
* mamba : use a state mask
It's cleaner than the previous heuristic of
checking for the pos of the first token in the batch.
inp_KQ_mask could not be re-used for this, because it has the wrong shape
and because it seems more suited to the next step of
simultaneous sequence processing (helping with the problem of
remembering which token belongs to which sequence(s)/state(s)).
* llama : replace the usage of n_ctx with kv_self.size in many places
* mamba : use n_tokens directly instead of n_tok
* mamba : in comments, properly refer to KV cells instead of slots
* mamba : reduce memory usage of ggml_ssm_scan
From 290.37 MiB to 140.68 MiB of CPU compute buffer size
with Mamba 3B with a batch size of 512.
The result tensor of ggml_ssm_scan was previously a big part
of the CPU compute buffer size. To make it smaller,
it does not contain the intermediate ssm states anymore.
Both y and the last ssm state are combined in the result tensor,
because it seems only a single tensor can be returned by an operator
with the way the graph is built.
* mamba : simultaneous sequence processing
A batch can now contain tokens from multiple sequences.
This is necessary for at least the parallel example, the server example,
and the HellaSwag test in the perplexity example.
However, for this to be useful, uses of llama_kv_cache_seq_rm/cp
will need to be changed to work on whole sequences.
* ggml : add ggml_ssm_conv as a new operator for the conv step of Mamba
This operator makes it possible to use and update the correct states
for each token of the batch in the same way as ggml_ssm_scan.
Other solutions which use existing operators would need loops which would
add too many nodes to the graph (at least the ones I thought of).
Using this operator further reduces the size of the CPU compute buffer
from 140.68 MiB to 103.20 MiB with Mamba 3B with a batch size of 512.
And (at least on CPU), it's a bit faster than before.
Note that "ggml_ssm_conv" is probably not the most appropriate name,
and it could be changed if a better one is found.
* llama : add inp_s_seq as a new input tensor
The most convenient implementation to select the correct state (for Mamba)
for each token is to directly get the correct index from a tensor.
This is why inp_s_seq is storing int32_t and not floats.
The other, less convenient way to select the correct state would be
to have inp_KQ_mask contain 1.0f for each state used by a token
and 0.0f otherwise. This complicates quickly fetching the first used
state of a token, and is also less efficient because a whole row
of the mask would always need to be read for each token.
Using indexes makes it easy to stop searching when there are
no more sequences for a token, and the first sequence assigned
is always very quickly available (it's the first element of each row).
* mamba : support llama_kv_cache_seq_cp copy chains
* mamba : support shifting and dividing the kv cache pos
* mamba : make the server and parallel examples work with whole sequences
A seq_id is dedicated to the system prompt in both cases.
* llama : make llama_kv_cache_seq_rm return whether it succeeded or not
* mamba : dedicate an input tensor for state copy indices
This is cleaner and makes it easier to adapt when/if token positions
(and by extension, inp_K_shift) are no longer integers.
* mamba : adapt perplexity, batched, and batched-bench examples
* perplexity : limit the max number of sequences
This adapts to what the loaded model can provide.
* llama : add llama_n_max_seq to get the upper limit for seq_ids
Used by the perplexity example.
* batched : pass n_parallel to the model's context params
This should have been there already, but it wasn't.
* batched-bench : reserve sequences to support Mamba
* batched-bench : fix tokens being put in wrong sequences
Generation quality isn't what's measured in there anyway,
but at least using the correct sequences avoids using non-consecutive
token positions.
* mamba : stop abusing attention metadata
This breaks existing converted-to-GGUF Mamba models,
but will allow supporting mixed architectures like MambaFormer
without needing to break Mamba models.
This will also allow changing the size of Mamba's states
without having to reconvert models in the future.
(e.g. using something else than d_conv - 1 columns for the conv_states
will not require breaking existing converted Mamba models again)
* gguf-py : add new KV metadata key-value pairs for Mamba
* llama : add new metadata key-value pairs for Mamba
* llama : guard against divisions by zero when n_head is 0
* mamba : rename "unlimited" KV cache property to "recurrent"
* mamba : more correctly update the "used" field of the KV cache
* ggml : in ggml_ssm_scan, use a threshold for soft_plus
This is how the official Mamba implementation does it,
and it's also what torch.nn.Softplus does.
* convert : for Mamba, fallback to internal NeoX tokenizer
The resulting models are exactly the same
as if the tokenizer.json and tokenizer_config.json of GPT-NeoX were there.
* mamba : support state saving and restoring
* ggml : implicitly pass src tensors through dst for Mamba-related ops
* mamba : clarify some comments
* server : fix cache_tokens not getting correctly resized
Otherwise, when the "we have to evaluate at least 1 token" special case
was triggered, an extra token was kept in cache_tokens even if it was
removed from the KV cache.
For Mamba, this caused useless prompt reprocessing when the previous
request triggered the above case.
* convert-hf : support new metadata keys for Mamba
For the models available at
https://huggingface.co/collections/state-spaces/transformers-compatible-mamba-65e7b40ab87e5297e45ae406
* mamba : rename metadata to be more similar to transformers library
This breaks existing converted-to-GGUF models,
but the metadata names are more "standard".
* mamba : support mamba-*-hf models
These models share their token_embd.weight with their output.weight
* mamba : add missing spaces
This is purely a formatting change.
* convert-hf : omit output.weight when identical with token_embd.weight
Only for Mamba for now, but it might be relevant for other models eventually.
Most Mamba models actually share these two tensors, albeit implicitly.
* readme : add Mamba to supported models, and add recent API changes
* mamba : move state_seq and state_mask views outside layer loop
A few tensors were also missing `struct` in front of `ggml_tensor`.
MSVC gives the following error with the existing macros:
`Error C2059 : syntax error: ','`
This patch adds `##` as a prefix to `__VA_ARGS__` to address this error.
* server : refactoring (wip)
* server : remove llava/clip objects from build
* server : fix empty prompt handling + all slots idle logic
* server : normalize id vars
* server : code style
* server : simplify model chat template validation
* server : code style
* server : minor
* llama : llama_chat_apply_template support null buf
* server : do not process embedding requests when disabled
* server : reorganize structs and enums + naming fixes
* server : merge oai.hpp in utils.hpp
* server : refactor system prompt update at start
* server : disable cached prompts with self-extend
* server : do not process more than n_batch tokens per iter
* server: tests: embeddings use a real embeddings model (#5908)
* server, tests : bump batch to fit 1 embedding prompt
* server: tests: embeddings fix build type Debug is randomly failing (#5911)
* server: tests: embeddings, use different KV Cache size
* server: tests: embeddings, fixed prompt do not exceed n_batch, increase embedding timeout, reduce number of concurrent embeddings
* server: tests: embeddings, no need to wait for server idle as it can timout
* server: refactor: clean up http code (#5912)
* server : avoid n_available var
ggml-ci
* server: refactor: better http codes
* server : simplify json parsing + add comment about t_last
* server : rename server structs
* server : allow to override FQDN in tests
ggml-ci
* server : add comments
---------
Co-authored-by: Pierrick Hymbert <pierrick.hymbert@gmail.com>
* fix mul_mat fault in cpy_f32_f16
* rm unused function
* add wait() for memcpy
* restore ci/run.sh, rename struct defination, fix bug in ggml_sycl_op_mul_mat_sycl
* fix format issue
* llama : fix segfault from unknown model arch name (#5820)
* llama : fix segfault from unknown model arch name
* llama : make all LLM maps const
This also requires using `std::map::at` instead of its `operator[]`
which does not exist for const maps.
* llama : name LLM_ARCH_UNKNOWN to "(unknown)"
This avoids errors from `std::map::at` when
getting the general name of the model architecture.
Using "(unknown)" instead of an empty string as per suggestion
https://github.com/ggerganov/llama.cpp/pull/5820#issuecomment-1973735284
* llama : remove redundant inner const for LLM_TENSOR_NAMES
The extra const won't do anything here as const maps
return const references to values.
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* llama : remove redundant nullptr check in llm_arch_from_string
Since LLM_ARCH_NAMES is a const map, no spurious elements
with a NULL name are inserted anymore, so this check is dead code.
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* llama : refactor internal quantization functions (#5830)
* scripts : add pod-llama.sh
* ggml : IQ3_S improvements (#5829)
* iq3_s: somewhat faster AVX2 dot product
On Ryzen a 7950X TG-128 increases to 16 t/s from 15.5 t/s using
16 threads. For 8 threads it is 13.85 t/s vs 11.75 t/s.
PP-512 increases to 28.5 t/s from 23.8 t/s.
* iq3_s: somewhat faster ARM_NEON dot product
Still dog slow - 10.7 t/s up from 9.9 t/s.
* iq3_s: another small ARM_NEON improvement
10.7 -> 11.0 t/s. Using vmulq_s8 is faster than the xor - sub trick
that works best on AVX2.
* iq3_s: minor improvement on Metal
49.4 t/s -> 50.3 t/s
* iq3_s: PPL improvement
E.g., for a context of 4096 LLaMA-v2-7B goes to 5.1340 from 5.1653.
* iq3_s: use new grid everywhere
* Fix ARM_NEON
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* convert-hf : make model class definitions self-contained (#5825)
* convert : automatically fall back to HfVocab if tokenizer.model doesn't exist (#5821)
* ggml : fix IQ3_S AVX implementation (#5834)
ggml-ci
* llama : add abort_callback to interrupt computation (#5409)
* using abort_callback from ggml to stop llama computation
* format fix
* a brief explaining comment
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: tests: passkey challenge / self-extend with context shift demo (#5832)
* server: tests: add models endpoint scenario
* server: /v1/models add some metadata
* server: tests: add debug field in context before scenario
* server: tests: download model from HF, add batch size
* server: tests: add passkey test
* server: tests: add group attention params
* server: do not truncate prompt tokens if self-extend through group attention is enabled
* server: logs: do not truncate log values
* server: tests - passkey - first good working value of nga
* server: tests: fix server timeout
* server: tests: fix passkey, add doc, fix regex content matching, fix timeout
* server: tests: fix regex content matching
* server: tests: schedule slow tests on master
* server: metrics: fix when no prompt processed
* server: tests: self-extend add llama-2-7B and Mixtral-8x7B-v0.1
* server: tests: increase timeout for completion
* server: tests: keep only the PHI-2 test
* server: tests: passkey add a negative test
* flake.lock: Update (#5842)
Flake lock file updates:
• Updated input 'flake-parts':
'github:hercules-ci/flake-parts/b253292d9c0a5ead9bc98c4e9a26c6312e27d69f' (2024-02-01)
→ 'github:hercules-ci/flake-parts/f7b3c975cf067e56e7cda6cb098ebe3fb4d74ca2' (2024-03-01)
• Updated input 'flake-parts/nixpkgs-lib':
'github:NixOS/nixpkgs/97b17f32362e475016f942bbdfda4a4a72a8a652?dir=lib' (2024-01-29)
→ 'github:NixOS/nixpkgs/1536926ef5621b09bba54035ae2bb6d806d72ac8?dir=lib' (2024-02-29)
• Updated input 'nixpkgs':
'github:NixOS/nixpkgs/cbc4211f0afffe6dfd2478a62615dd5175a13f9a' (2024-02-23)
→ 'github:NixOS/nixpkgs/1536926ef5621b09bba54035ae2bb6d806d72ac8' (2024-02-29)
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* server : init http requests thread pool with --parallel if set (#5836)
* ci : schedule slow server tests only on Release or on demand (#5839)
* llama : fix llama_copy_state_data with fragmented KV cache (#5840)
The row size of the saved states was based on kv_self.head while
it should be based on llama_kv_cache_cell_max.
Existing session files should still work.
* llama : fix llama_kv_cache_cell_max inability to return 1
I've also changed its return type to uint32_t,
because this function is always used to set the value of uint32_t variables,
and because the index already has this type.
* llama : fix state size calculation
Some bytes in the state were unaccounted for in llama_get_state_size.
Since the logits reserve so much space, it did not cause problems.
* gguf-dump : support i-quants (#5841)
Co-authored-by: Black_Fox <radekliska@gmail.com>
* llama : allow for user specified embedding pooling type (#5849)
* allow for user specified pooling type
* llama : use enum types over int
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* readme : add API changes section
* cuda : fix data race in soft max (#5853)
* main : support special tokens as reverse/anti prompt (#5847)
* Support special tokens as reverse/anti prompt.
* Tokenize antiprompts only once.
* main : minor
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* common : use LLAMA_DEFAULT_SEED (#5855)
* add some new ops, fix some operators and add batch operations to certain operators. (ggml/747)
* cuda: fix group_norm
* cuda: add batch inference support for ggml_pad/ggml_upscale
* add ggml_arrange
* add ggml_timestep_embedding
* update ggml_arange/ggml_timestep_embedding tests
* cuda: fix im2col
* add ggml_arange/ggml_timestep_embbeding support for metal backend
* fix some bugs
* fix some bugs
* Update ggml.h
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml-cuda.cu
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml-metal.m
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml-metal.m
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml-metal.metal
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* modify according to the review comments
* ggml : fix compile warnings + code style
* ggml : normalize compute_forward calls + fix seg fault in debug
* minor
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
* sync : ggml
* add alias for chat template (#5858)
* speculative : implement stochastic speculative sampling (#5625)
* (WIP) Implement stochastic speculative decoding
* sample from residual distribution on draft accept failure
* fix#5657: force greedy sampling with probs when temp is 0
* remove p_accept parameter
* fix style
* remove unused variables
* add srand() in speculative.cpp
* replace use of rand() with mt19937 sampling
* fixes based on review (@JohannesGaessler)
* fix r random generation
* randomly select next sequence to verify + fix bug in memory freeing
* fix bug in active_seqs sync
* fix uniform int distribution initialization
* remove warnings from comparison between int and size_t
* check grammar in `llama_sample_probability_distribution_impl`
* remove malloc code by utilizing vectors
* add PR link to README
* cmake : handle cases where git index is not found in .git (#5844)
* Update CMakeLists.txt
* Update CMakeLists.txt
* ggml : introduce ggml_status (ggml/750)
* using enum as an exit code instead of macros
* update return type from enum to unsigned int
* indentation fix
* compound update
ggml_compute_exit_code -> ggml_status
changed ggml_status from a bit-field type to simple codes
ggml_status to string cast
* ggml_status to string cast
* GGML_CALL was removed
Co-authored-by: slaren <slarengh@gmail.com>
---------
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* sync : ggml
ggml-ci
* ggml : fix unknown status (#0)
* flake : fix
* llama : fix embeddings (#5796)
* llama : fix embeddings
ggml-ci
* llama : do not use KV cache for non-causal models
ggml-ci
* embeddings : fix llama_batch_init arg
* llama : add pooling switch
* llama : distinguish token vs sequence embeddings
ggml-ci
* llama : assert pooling tensor
* llama : simplify causal mask condition
ggml-ci
* llama : assert input batch with pooling enabled
* readme : update API changes list
* nix: static build (#5814)
* fix speculative decoding build on windows (#5874)
* rebase and rm tailing space
---------
Co-authored-by: LiangtaoJin <liang-tao.jin@intel.com>
Co-authored-by: compilade <113953597+compilade@users.noreply.github.com>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Kawrakow <48489457+ikawrakow@users.noreply.github.com>
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
Co-authored-by: Michael Podvitskiy <podvitskiymichael@gmail.com>
Co-authored-by: Pierrick Hymbert <pierrick.hymbert@gmail.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Nindaleth <Nindaleth@users.noreply.github.com>
Co-authored-by: Black_Fox <radekliska@gmail.com>
Co-authored-by: Douglas Hanley <thesecretaryofwar@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: DAN™ <dranger003@gmail.com>
Co-authored-by: leejet <leejet714@gmail.com>
Co-authored-by: Minsoo Cheong <54794500+mscheong01@users.noreply.github.com>
Co-authored-by: Dane Madsen <dane_madsen@hotmail.com>
Co-authored-by: hutli <6594598+hutli@users.noreply.github.com>
Co-authored-by: Jeffrey Quesnelle <emozilla@nousresearch.com>
* using enum as an exit code instead of macros
* update return type from enum to unsigned int
* indentation fix
* compound update
ggml_compute_exit_code -> ggml_status
changed ggml_status from a bit-field type to simple codes
ggml_status to string cast
* ggml_status to string cast
* GGML_CALL was removed
Co-authored-by: slaren <slarengh@gmail.com>
---------
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* (WIP) Implement stochastic speculative decoding
* sample from residual distribution on draft accept failure
* fix#5657: force greedy sampling with probs when temp is 0
* remove p_accept parameter
* fix style
* remove unused variables
* add srand() in speculative.cpp
* replace use of rand() with mt19937 sampling
* fixes based on review (@JohannesGaessler)
* fix r random generation
* randomly select next sequence to verify + fix bug in memory freeing
* fix bug in active_seqs sync
* fix uniform int distribution initialization
* remove warnings from comparison between int and size_t
* check grammar in `llama_sample_probability_distribution_impl`
* remove malloc code by utilizing vectors
* add PR link to README
* Support special tokens as reverse/anti prompt.
* Tokenize antiprompts only once.
* main : minor
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
The row size of the saved states was based on kv_self.head while
it should be based on llama_kv_cache_cell_max.
Existing session files should still work.
* llama : fix llama_kv_cache_cell_max inability to return 1
I've also changed its return type to uint32_t,
because this function is always used to set the value of uint32_t variables,
and because the index already has this type.
* llama : fix state size calculation
Some bytes in the state were unaccounted for in llama_get_state_size.
Since the logits reserve so much space, it did not cause problems.
* server: tests: add models endpoint scenario
* server: /v1/models add some metadata
* server: tests: add debug field in context before scenario
* server: tests: download model from HF, add batch size
* server: tests: add passkey test
* server: tests: add group attention params
* server: do not truncate prompt tokens if self-extend through group attention is enabled
* server: logs: do not truncate log values
* server: tests - passkey - first good working value of nga
* server: tests: fix server timeout
* server: tests: fix passkey, add doc, fix regex content matching, fix timeout
* server: tests: fix regex content matching
* server: tests: schedule slow tests on master
* server: metrics: fix when no prompt processed
* server: tests: self-extend add llama-2-7B and Mixtral-8x7B-v0.1
* server: tests: increase timeout for completion
* server: tests: keep only the PHI-2 test
* server: tests: passkey add a negative test
* using abort_callback from ggml to stop llama computation
* format fix
* a brief explaining comment
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* iq3_s: somewhat faster AVX2 dot product
On Ryzen a 7950X TG-128 increases to 16 t/s from 15.5 t/s using
16 threads. For 8 threads it is 13.85 t/s vs 11.75 t/s.
PP-512 increases to 28.5 t/s from 23.8 t/s.
* iq3_s: somewhat faster ARM_NEON dot product
Still dog slow - 10.7 t/s up from 9.9 t/s.
* iq3_s: another small ARM_NEON improvement
10.7 -> 11.0 t/s. Using vmulq_s8 is faster than the xor - sub trick
that works best on AVX2.
* iq3_s: minor improvement on Metal
49.4 t/s -> 50.3 t/s
* iq3_s: PPL improvement
E.g., for a context of 4096 LLaMA-v2-7B goes to 5.1340 from 5.1653.
* iq3_s: use new grid everywhere
* Fix ARM_NEON
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* llama : fix segfault from unknown model arch name
* llama : make all LLM maps const
This also requires using `std::map::at` instead of its `operator[]`
which does not exist for const maps.
* llama : name LLM_ARCH_UNKNOWN to "(unknown)"
This avoids errors from `std::map::at` when
getting the general name of the model architecture.
Using "(unknown)" instead of an empty string as per suggestion
https://github.com/ggerganov/llama.cpp/pull/5820#issuecomment-1973735284
* llama : remove redundant inner const for LLM_TENSOR_NAMES
The extra const won't do anything here as const maps
return const references to values.
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* llama : remove redundant nullptr check in llm_arch_from_string
Since LLM_ARCH_NAMES is a const map, no spurious elements
with a NULL name are inserted anymore, so this check is dead code.
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* suport multiple cards: split-mode - layer|row
* rm warning
* rebase with master, support tow new OPs, close feature for -sm=row, fix for unit test
* update news
* fix merge error
* update according to review comments
Reduces peak tmpfs usage and should prevent the check from failing from
running out of space.
Fixes the 'No space left on device' issue mentioned in #5703.
* add magika inference example
* ggml : fix unaligned accesses in custom ops
* ggml : fix FP32 GELU for values that exceed the FP16 range
* use ggml_pool_1d
* add README
* Update README.md
* pad inputs if the files are too small
* cleanup
ggml-ci
* Introduce backend GUIDs
Initial proposed implementation of backend GUIDs
(Discussed in https://github.com/ggerganov/ggml/pull/741)
Hardcoded CPU backend GUID (for now)
Change ggml_backend_is_cpu logic to use GUID
* Remove redundant functions
Remove redundant functions `ggml_backend_i::get_name` and `ggml_backend_guid` which are not desired for future expansion
* Add spaces to match style
Co-authored-by: slaren <slarengh@gmail.com>
* Fix brace style to match
Co-authored-by: slaren <slarengh@gmail.com>
* Add void to () in function signature
Co-authored-by: slaren <slarengh@gmail.com>
* Add back ggml_backend_guid and make CPU_GUID a local static in ggml_backend_cpu_guid
* add guids to all backends
ggml-ci
---------
Co-authored-by: slaren <slarengh@gmail.com>
* implement nfd for stripping accents in wpm tokenizer
* sort nfd map; reuse iterator
* use builtin tolower
* add locale include
* Simplify to_lower cases
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Add "/chat/completions" as alias for "/v1/chat/completions"
* merge to upstream master
* minor : fix trailing whitespace
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* WIP: make i-quants work for QK_K = 64
* iq2_xs: attempt to fix AVX dot product for QK_K = 64
Tests pass, but I get gibberish.
* QK_K = 64 tests pass on ARM_NEON and Metal
Sadly, that does not mean it actually works.
* Make CUDA compile with QK_K = 64
Tests don't pass, plus we get misaligned access
* Q2_K: fixed bug in imatrix quantization for QK_K = 64
* iq1_s: turn off SIMD implementation for QK_K = 64 (it does not work)
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Try IQ4_NL with blocks of 64 - does not look good
* iq4_xs: go to super-blocks of 256 and 6-bit scales for blocks of 32
* iq4_xs: CUDA works - 133.2 t/s
* iq4_xs: AVX2 dot product
* iq4_xs: ARM_NEON dot product
* iq4_nl: Metal implementation
As usual, Metal / Apple Silicon don't like my quants.
* iq3_xs: minor fix
* iq4_xs: shrink by using IQ3_S for attn_k and attn_q
* iq4_xs: revert using IQ3_S for attn_k and attn_v
PPL vs size is good, but CPU performance suffers: on M2 Max
TG-128 drops to 21.7 t/s from 28.8, and on a Ryzen-7950X
to 14.5 t/s from 15.8 t/s. On CUDA we have 135 t/s when
using IQ3_S vs 133 t/s with pure IQ4_XS.
* Fix CI
* iq4_xs: Added forgotten check for 256 divisibility
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding IQ2_S and IQ2_M as a single cumulative commit
* Update examples/quantize/quantize.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: docs - refresh and tease a little bit more the http server
* Rephrase README.md server doc
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/server/README.md
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/server/README.md
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update README.md
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
The system prompt is now decoded in batches.
* server : fix off-by-one n_past when start of prompt matches whole cache
The tokens right after the matching part would otherwise skip a pos value.
* [ggml-quants] Provide ggml_vqtbl1q_u8 for 64bit compatibility
vqtbl1q_u8 is not part of arm v7 neon library
* [android-example] Remove abi filter after arm v7a fix
* [github-workflows] Do not skip Android armeabi-v7a build
* server: logs - always use JSON logger, add add thread_id in message, log task_id and slot_id
* server : skip GH copilot requests from logging
* server : change message format of server_log()
* server : no need to repeat log in comment
* server : log style consistency
* server : fix compile warning
* server : fix tests regex patterns on M2 Ultra
* server: logs: PR feedback on log level
* server: logs: allow to choose log format in json or plain text
* server: tests: output server logs in text
* server: logs switch init logs to server logs macro
* server: logs ensure value json value does not raised error
* server: logs reduce level VERBOSE to VERB to max 4 chars
* server: logs lower case as other log messages
* server: logs avoid static in general
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: logs PR feedback: change text log format to: LEVEL [function_name] message | additional=data
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: monitoring - add /metrics prometheus compatible endpoint
* server: concurrency issue, when 2 task are waiting for results, only one call thread is notified
* server: metrics - move to a dedicated struct
* iq4_nl: squash commits for easier rebase
* Basics (quantize, dequantize)
* CUDA dequantize and dot product
* Slightly faster CUDA dot product (120 t/s)
* Switch to 6-bit scales
* Scalar dot product
* AVX2 dot product
* ARM_NEON dot product
* Works on metal, but still slow
* Slightly better Metal dot product
* Another small Metal improvement
* Metal dot product is getting there
* Faster CUDA dot product
* Add 1/8 ffn_down layers as Q5_K when no imatrix has been provided
* Report the actual bpw
* Add _xs mix that is 4.05 bpw for non-MoE models
* Remove IQ4_XS for now, slightly adjust kvalues_iq4nl
* AVX2 dot product uses Q8_0 instead of Q8_K
* Add to test-backend-ops
* Minor fix
* Also use use Q5_K for attn_output in MoE models
* Fixes after merging latest master
* Switching to blocks of 32
* AVX2 for blocks of 32
* Scaler dot product for blocks of 32
* ARM_NEON dot product for blocks of 32
* Metal kernels for blocks of 32
* Slightly faster Metal kernels
* Resurrecting iq3_xs
After all the experimentation, nothing was better than this.
* Minor PPL improvement via a block scale fudge factor
* Minor improvement via 3 neighbours
* iq3_xs: working scalar and AVX2 dot products
* iq3_xs: ARM_NEON dot product - works but extremely slow (10 t/s)
* iq3_xs: working Metal implementation
* Adding IQ3_M - IQ3_XS mix with mostly Q4_K
* iiq3_xs: a 3.4375 bpw variant
* iq3_xs: make CUDA work for new version
* iq3_xs: make scalar and AVX2 work for new version
* iq3_s: make ARM_NEON work with new version
* iq3_xs: make new version work on metal
Performance is very similar to Q3_K_S
* iq3_xs: tiny Metal speed improvement
* iq3_xs: tiny Metal speed improvement
* Fix stupid warning
* Q3_K_XS now uses a mix of IQ3_XS and IQ3_XXS
* iq3_xs: rename to iq3_s
* iq3_s: make tests pass
* Move Q3_K_XS mix to 3.25 bpw
* Attempt to fix failing tests
* Another attempt to fix the Windows builds
* Attempt to fix ROCm
* ROCm again
* iq3_s: partial fix for QK_K = 64
* iq3_s: make it work on metal for QK_K = 64
Pleasent surprise: the coding was super-block size independent,
so all it took was to delete some QK_K == 256 guards.
* Will this fix ROCm?
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* server: tests: init scenarios
- health and slots endpoints
- completion endpoint
- OAI compatible chat completion requests w/ and without streaming
- completion multi users scenario
- multi users scenario on OAI compatible endpoint with streaming
- multi users with total number of tokens to predict exceeds the KV Cache size
- server wrong usage scenario, like in Infinite loop of "context shift" #3969
- slots shifting
- continuous batching
- embeddings endpoint
- multi users embedding endpoint: Segmentation fault #5655
- OpenAI-compatible embeddings API
- tokenize endpoint
- CORS and api key scenario
* server: CI GitHub workflow
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Exposes a few attributes demonstrating how to build [singularity](https://docs.sylabs.io/guides/latest/user-guide/)/[apptainer](https://apptainer.org/) and Docker images re-using llama.cpp's Nix expression.
Built locally on `x86_64-linux` with `nix build github:someoneserge/llama.cpp/feat/nix/images#llamaPackages.{docker,docker-min,sif,llama-cpp}` and it's fast and effective.
GitHub does not expose environment and repository variables to PRs coming from forks implies that we've been disabling the Nix CI actions for most PRs.
The `if:` also didn't make much sense, because we can always pull from cachix, and there's no point (albeit no risk either) in pushing cache for the untrusted code.
* server: fallback to chatml
* add new chat template
* server: add AlphaMonarch to test chat template
* server: only check model template if there is no custom tmpl
* remove TODO
* server: health: fix race condition on slots data using tasks queue
* server: health:
* include_slots only if slots_endpoint
* fix compile warning task.target_id not initialized.
This commit adds the `--skip-unknown` option to the convert.py script
and removes the saving of the updated checkpoints to avoid updating
possibly checked out files.
The motivation for this change is that this was done for 1.5
in Commit fc0c8d286a ("llava :
update surgery script to not remove tensors") and makes the examples
more consistent.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
There are couple things in this architecture:
1. Shared input and output embedding parameters.
2. Key length and value length are not derived from `n_embd`.
More information about the models can be found at
https://ai.google.dev/gemma. GGUFs can be downloaded from
https://huggingface.co/google.
* iq4_nl: squash commits for easier rebase
* Basics (quantize, dequantize)
* CUDA dequantize and dot product
* Slightly faster CUDA dot product (120 t/s)
* Switch to 6-bit scales
* Scalar dot product
* AVX2 dot product
* ARM_NEON dot product
* Works on metal, but still slow
* Slightly better Metal dot product
* Another small Metal improvement
* Metal dot product is getting there
* Faster CUDA dot product
* Add 1/8 ffn_down layers as Q5_K when no imatrix has been provided
* Report the actual bpw
* Add _xs mix that is 4.05 bpw for non-MoE models
* Remove IQ4_XS for now, slightly adjust kvalues_iq4nl
* AVX2 dot product uses Q8_0 instead of Q8_K
* Add to test-backend-ops
* Minor fix
* Also use use Q5_K for attn_output in MoE models
* Fixes after merging latest master
* Switching to blocks of 32
* AVX2 for blocks of 32
* Scaler dot product for blocks of 32
* ARM_NEON dot product for blocks of 32
* Metal kernels for blocks of 32
* Slightly faster Metal kernels
* iq4_nl: Fix after merging with master
* iq4_nl: another fix after merging with master
* Use IQ4_NL instead of Q4_K when using k-quants is not possible
* Fix typo that makes several tests fail
* It was the ggml_vdotq thing missed inside the brackets
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit contains a suggestion for the README.md in the llava
example. The suggestion adds explicit instructions for how to convert
a llava-1.6 model and run it using llava-cli.
The motivation for this is that having explicit instructions similar to
the 1.5 instructions will make it easier for users to try this out.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* add build support for embedded metal library
* Update Makefile
---------
Co-authored-by: Haoxiang Fei <feihaoxiang@idea.edu.cn>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* support minLength and maxLength in JSON schema grammar converter
* Update examples/json-schema-to-grammar.py
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ggml : embed Metal library source (ggml-metal.metal) into binary
enable by setting WHISPER_EMBED_METAL_LIBRARY
* rename the build option
* rename the preprocessor directive
* generate Metal library embedding assembly on-fly during build process
This is a follup of Commit fc0c8d286a
("llava : update surgery script to not remove tensors") but this time
the change is to the BakLLaVA specific part of the surgery script.
I've been able to test this using SkunkworksAI/BakLLaVA-1 and it works
as expected using the instructions in README.md.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* llama: add llama_chat_apply_template
* test-chat-template: remove dedundant vector
* chat_template: do not use std::string for buffer
* add clarification for llama_chat_apply_template
* llama_chat_apply_template: add zephyr template
* llama_chat_apply_template: correct docs
* llama_chat_apply_template: use term "chat" everywhere
* llama_chat_apply_template: change variable name to "tmpl"
* #ifdef out some code NUMA blocks for Android due to lack of support
* added in some __ANDROID__ if def gates around numa code and forced GLIBC prior to 2.29 to use a syscall for getcpu instead of the wrapper
* Changed gates on numa platform specific stuff to __gnu_linux__ to skip any platforms without glibc
* harmonizing #if defined blocks for numa code to __gnu_linux__ since that's the only model that's being followed anyways
---------
Co-authored-by: root <root@nenya.lothlorien.ca>
* build : pass all warning flags to nvcc via -Xcompiler
* make : fix apparent mis-merge from #3952
* make : fix incorrect GF_CC_VER for CUDA host compiler
* server: enrich health endpoint with available slots, return 503 if not slots are available
* server: document new status no slot available in the README.md
This commit updates the surgery script to not remove the tensors from the
model file. For this to work the `--skip-unknown` flag is added as an
argument to the convert.py script in README.md.
The motivation for this change is that the surgery script currently
removes the projector tensors from the model file. If the model was
checked out from a repository, the model file will have been updated
and have to be checked out again to reset this effect. If this can be
avoided I think it would be preferable.
I did not perform this change for BakLLaVA models as I am not sure
how that part works.
* iq1_s: WIP basics
* iq1_s: CUDA is working
* iq1_s: scalar CPU dot product
* iq1_s: WIP AVX2 dot product - something is not right
* Fix tests
* Fix shadow warnings
* Fix after merge with latest master
* iq1_s: AVX2 finally works
* iq1_s: ARM_NEON dot product. Works, but not very fast
* iq1_s: better grid
* iq1_s: use IQ2_XXS for attn_output
At a cost of 0.04 extra bpw this gives a big improvement in PPL.
* iq1_s: Metal basics
Dequantize works, but not dot product
* iq1_s: Metal works, but quite slow
As usual, Apple Silicon does not like the code I write.
* iq1_s: Tests
* iq1_s: slightly faster dot product
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* ggml : avoid recomputing alibi slopes (CPU)
* llama : reuse hparams.f_max_alibi_bias in all cases
ggml-ci
* ggml : support alibi bias in ggml_soft_max_ext (CPU + Metal)
ggml-ci
* ggml : handle all SRCs (do not break on first null)
ggml-ci
* tests : do not use slope for large soft_max
accumulates too much error
ggml-ci
* ggml : alternative ALiBi without extra tensor
We compute the slopes in the kernel
ggml-ci
* cuda : add ALiBi support in ggml_soft_max_ext
ggml-ci
* ggml : deprecate ggml_alibi
* ggml : support multi-sequence ALiBi (Metal)
ggml-ci
* cuda : add multi-seq ALiBi + remote F16 soft_max
ggml-ci
* ggml : update deprecation message
* ggml : fix pos ptr when no ALiBi
ggml-ci
* cuda : fix performance (pow -> powf)
* cuda : precompute ALiBi constants
* metal : pre-compute ALiBi slopes
ggml-ci
* llama : init kq_pos only if needed
ggml-ci
* test-backend-ops : add null pos test to soft_max
test-backend-ops : replace soft_max tests
ggml-ci
---------
Co-authored-by: slaren <slarengh@gmail.com>
* scripts : add helpers script for bench comparing commits
* scripts : detect CUDA
* set flags after checking the command line
* fix make flags
---------
Co-authored-by: slaren <slarengh@gmail.com>
* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h
* Reverted Makefile
* Fixed include
* Removed sched.h from ggml.h, moved ggml_get_numa_affinity into ggml.c, removed trailing whitespace and fixed up a few inconsistent variables
* removed trailing whitespace
* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h
* Reverting Makefile
* Fixed a number of issues with the move from BOOL to ggml_numa_strategies. Added a note about mirror mode note being implemented yet
* Removing MIRROR_MODE code for this PR
* Removing last bit of MIRROR_MODE code for this PR
* Removing unneeded branch in server.cpp example and moving get_numa_affinity and making it static
* Fixed lingering init_llama_backend() bool calls in tests and examples
* Remote enum llama_numa_strategies
* Revert bad merge with dynatemp flags
* add missing enum ggml_numa_strategies declaration and revert sync problem with master
* add missing enum ggml_numa_strategies declaration
* fixed ggml_init_numa variable
* Update ggml.h
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update READMEs with info about numa flags, change INTERLEAVE strategy name to DISTRIBUTE everywhere, implement the improved distribution strategy from @rankaiyx, fix a spelling mistake and un-merge some bad merges
* split numa init out from llama_backend_init and created llama_numa_init. Updated all code paths and samples
* Fix up some boolean vs enum comparisons
* Added #ifdefs for non-Linux OS that don't have cpu_set_t datatype
* Update ggml.h
Align enum values
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml.c
Remove whitespace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml.c
align paremeters
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/server/server.cpp
remove whitespace and align brace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update common/common.cpp
Remove whitespace and align brace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* unified ggml_numa_strategy enum and fixed text alignment in server.cpp example
* Update ggml.c
simplified return for platforms without NUMA support
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* removed redundant else from cli argument processing of --numa
* whitespace
---------
Co-authored-by: root <root@nenya.lothlorien.ca>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
* llava: fix clip-model-is-vision flag in README.md
This commit fixes the flag `--clip_model_is_vision` in README.md which
is does not match the actual flag:
```console
$ python convert-image-encoder-to-gguf.py --help
...
--clip-model-is-vision
The clip model is a pure vision model
(ShareGPT4V vision extract for example)
```
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* llava: update link to vit config in README.md
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* fix(gguf-py): special tokens are no longer skipped when add_<token>_token is set to false
* fix(gguf-py): added missing cls and mask token ids to the gguf metadata
* Fix memory management in llava and server code
Fixes this error:
llama_new_context_with_model: graph splits (measure): 3
Available slots:
-> Slot 0 - max context: 6000
{"timestamp":1707926446,"level":"INFO","function":"main","line":2623,"message":"model loaded"}
all slots are idle and system prompt is empty, clear the KV cache
slot 0 - loaded image
slot 0 is processing [task id: 0]
slot 0 : kv cache rm - [0, end)
slot 0 - encoding image [id: 1]
munmap_chunk(): invalid pointer
Aborted
* Make it cleaner by checking size in batch free wrapper
* Create llava-survery-v2.py
* Update convert-image-encoder-to-gguf.py
* Update convert-image-encoder-to-gguf.py
* Rename llava-survery-v2.py to llava-surgery-v2.py
* Update convert-image-encoder-to-gguf.py
will now search for projector
* Update convert-image-encoder-to-gguf.py
whoops
* Update llava-surgery-v2.py
* Clip: Bugfix for normalization (it did not loat the 3 std and mean values)
Clip: bicubic resize function
Clip: added save-to-bmp/pil for debugging and conversion from/to 32/8 images
Clip: added normalization with FP16 precision simulation (image tensors match HF implementation, can be switched off, only used for llava-1.6)
Clip: added newline tensor, mergetype kv, image-grid kv, new resize-pad function with resolution from gridpoints
Clip: clip_image_preprocess now returns a float * vector instead of float, this way llava 1.5 and 1.6 is supported
llava: added ggml cpu graph for embedding patching, added spatial_unpad preliminary support, added a lot of comments that need to be cleaned when all is final
convert-image-encoder: fixed image-grid flattening
* whitespace corrections
* ws
* Tensors are now properly permuted.
Before the embeddings were inserted 1:1, now they are split into the 24x24 patches as in reference.
* ws
* added verbose_prompt support into cli
added stopwords for llava-1.6 into cli
* moved llava functions to llava.cpp, made clip.h C compatible API, replaced vector style functions with pointers, added a debug define to remove functions from compilation while not needed
* ws
* convert : skip unknown tensors (need for LLaVA)
* llava : update readme
* llava : fix compile warnings
* llava : style
* convert : add --skip-unknown CLI arg
* server : remove clip structs
* bugfix for non llava-1.6
It should now work with llava-1.5 as well
* clip : minor code rearrange
* llava : update readme a bit
---------
Co-authored-by: John <cmt-nct@users.noreply.github.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Early return for zero size calls to get_tensor.
Signed-off-by: Adam Treat <treat.adam@gmail.com>
* Update ggml-kompute.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml-kompute.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Add an early return to the get/set tensor when the size is null.
Signed-off-by: Adam Treat <treat.adam@gmail.com>
* Early return after the assertions.
Signed-off-by: Adam Treat <treat.adam@gmail.com>
* Since we do the early return in the generic backend now no reason to do so here as well.
Signed-off-by: Adam Treat <treat.adam@gmail.com>
---------
Signed-off-by: Adam Treat <treat.adam@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* finetune: rename feed-forward tensors (w1/w2/w3)
This commit renames the feed-forward tensors w1, w2 and w3 to ffn_gate,
ffn_down and ffn_up respectively.
The motivation for this change is to make it easier to understand the
purpose of the tensors. This also seems to be inline with the names
used in the llama_layer struct in llama.cpp.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* train-text-from-scratch: rename ff tensors
This commit renames the feed-forward tensors w1, w2 and w3 to ffn_gate,
ffn_down and ffn_up respectively.
The motivation for this change is to make it easier to understand the
purpose of the tensors. This also seems to be inline with the names
used in the llama_layer struct in llama.cpp
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* batched embedding: pool outputs by sequence id. updated embedding example
* bring back non-causal attention
* embd : minor improvements
* llama : minor
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* make: add error message for bad CUDA version
* Update Makefile
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* llava: remove prog parameter from ArgumentParser
This commit removes the `prog` parameter from `ArgumentParser`
so that it uses the default value which is the name of the script.
The motivation for this change is that currently the usage output looks
like this:
```console
$ python examples/llava/convert-image-encoder-to-gguf.py --help
usage: convert_hf_to_gguf.py [-h] ...
```
And with this change it will look like this:
```console
$ python examples/llava/convert-image-encoder-to-gguf.py --help
usage: convert-image-encoder-to-gguf.py [-h] ...
```
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* ci: add W503 to flake8 ignore list
This commit adds W503 to the ignore list for flake8. This is done to
avoid the following error:
W503 line break before binary operator
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* BERT model graph construction (build_bert)
* WordPiece tokenizer (llm_tokenize_wpm)
* Add flag for non-causal attention models
* Allow for models that only output embeddings
* Support conversion of BERT models to GGUF
* Based on prior work by @xyzhang626 and @skeskinen
---------
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* vulkan: refactor guess_matmul_pipeline for vendor
Refactor ggml_vk_guess_matmul_pipeline to simplify adding per-vendor
conditionals.
Signed-off-by: Sergio Lopez <slp@redhat.com>
* vulkan: only use M-sized matmul on Apple GPUs
L-sized and S-sized matmuls are broken on Apple GPUs, force using
M-size with this vendor.
Signed-off-by: Sergio Lopez <slp@redhat.com>
---------
Signed-off-by: Sergio Lopez <slp@redhat.com>
* common: use enums for sampler types
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* minor : spaces
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server: allow to specify tokens as strings in logit_bias
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ggml: aarch64: implement smmla kernel for q8_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q8_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_1_q8_1 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_1_q8_1 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: update unit tests for the new vec_dot interface
* llama.cpp: add MATMUL_INT8 capability to system_info
A common default for the maximum number of open files is 256, which can
lead to `asyncio.gather(*tasks)` failing with Too many open files.
$ python ggml_vk_generate_shaders.py --glslc=$ANDROID_NDK_PATH/shader-tools/darwin-x86_64/glslc
ggml_vulkan: Generating and compiling shaders to SPIR-V
Traceback (most recent call last):
File "/Users/neuman/Code.noindex/github/llama.cpp/ggml_vk_generate_shaders.py", line 2326, in <module>
asyncio.run(main())
File "/Users/neuman/Code.noindex/miniforge3/lib/python3.10/asyncio/runners.py", line 44, in run
return loop.run_until_complete(main)
File "/Users/neuman/Code.noindex/miniforge3/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
return future.result()
File "/Users/neuman/Code.noindex/github/llama.cpp/ggml_vk_generate_shaders.py", line 2294, in main
await asyncio.gather(*tasks)
[...snip...]
OSError: [Errno 24] Too many open files
This change sets a reasonable concurrency limit for tasks (and therefore
open files), without significant impact on run time.
* llava: add requirements.txt and update README.md
This commit adds a `requirements.txt` file to the `examples/llava`
directory. This file contains the required Python packages to run the
scripts in the `examples/llava` directory.
The motivation of this to make it easier for users to run the scripts in
`examples/llava`. This will avoid users from having to possibly run into
missing package issues if the packages are not installed on their system.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* llava: fix typo in llava-surgery.py output
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
This commit adds the missing .py extension to the convert-image-encoder-to-gguf
script. It also fixes the paths for the `model` and `mmproj` options in the
example llava-cli command.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* fix bug for norm_rms_eps missing
* to align with the same order as convert.py for model write
* fix: undo HF models permute tensor
* update for flake8 lint
This commit fixes a typo in the README.md file for the llava example
which is causing the formatting to look a little off:
Clone llava-v15-7b`` and clip-vit-large-patch14-336`` locally
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Initial Vulkan multi-gpu implementation
Move most global variables into backend context
* Add names to backend device functions
* Add further missing cleanup code
* Reduce code duplication in tensor split layer assignment
* generalize LLAMA_SPLIT_LAYER for all backends, do not expose device count and memory in llama.h
* Only do device info print in the beginning and initialize one backend for cpu assist
Add missing cleanup code
* Rework backend memory management to make sure devices and buffers get properly allocated and freed
* Rename cpu assist free function
---------
Co-authored-by: slaren <slarengh@gmail.com>
* support minicpm arch.
* fix tab/space typo.
* convert minicpm model via convert-hf-gguf.py
* try to make tokenizer work
* fix bug for quantize minicpm
* fix for flake8 lint
* remove convert-minicpm.py
* fix for editorconfig
* correct minicpm model type (size)
* constants expanded for minicpm
* Minor change of the constant names for minicpm
* include total "num_slots" in default_generation_settings_for_props
* cleanup total_slots return value in /props endpoint
* update /props endpoint docs with total_slots
* remove num_slots from default_generation_settings_for_props
* update /props endpoint section
server : fix deadlock when prompt array contains strings and numbers
server : removed an unnecessary generation when generating multi-prompts
server : removed an unnecessary assert
* Make use of ggml-quants.h possible in C++ code
* One cannot possibly be defining static_assert in a C++ compilation
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Avoid duplicating function calls when using MIN/MAX macros.
Since these copy "a" and "b" they ask the compiler to evaluate one of them twice. The compiler doesn't have a problem with removing the duplication in something like MAX(0, x + 2), but in some cases we're calling functions, and those calls just happen twice.
By explicitly evaluating at the expression we get smaller and faster code without duplicate calls. See ggml_rope_yarn_corr_dims in Compiler Explorer:
https://godbolt.org/z/Ee4KMrvKh
Code behaves exactly the same.
* Update ggml.c
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
We get slightly better PPL, and we cut quantization time in
nearly half.
The trick is to 1st quantize without forcing points onto the E8-lattice.
We can then use a narrower search range around the block scale that we
got that way.
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Update server-llm.sh
Add flag --non-interactive that allows run script without asking a permission
* Update scripts/server-llm.sh
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* imatrix: adding --combine and --continue-from
* imatrix: be able to start from a specific chunk
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Fix Vulkan on Intel ARC
Optimize matmul for Intel ARC
Add Vulkan dequant test
* Add Vulkan debug and validate flags to Make and CMakeLists.txt
* Enable asynchronous transfers in Vulkan backend
* Fix flake8
* Disable Vulkan async backend functions for now
* Also add Vulkan run tests command to Makefile and CMakeLists.txt
* update guide for make installation, memory, gguf model link, rm todo for windows build
* add vs install requirement
* update for gpu device check
* update help of llama-bench
* fix grammer issues
The llama_batch_init allocates memory for a fixed number of tokens.
However, the llama_batch_free only frees memory for the number of
tokens that were added to the batch.
This change-set uses a null terminated array for the batch seq_id, and
frees all the elements until the nullptr is reached. This change-set
also changes the name of the first parameter from `n_tokens` to
`n_tokens_alloc` to more clearly indicate that this value is the number
of tokens allocated to the batch, not the number of tokens in the batch.
* add --no-mmap, show sycl backend
* fix conflict
* fix code format, change print for --no-mmap
* ren no_mmap to mmap, show mmap when not default value in printer
* update guide for mmap
* mv position to reduce model reload
* New Feature:
1. Sum_Rows:
fix cuda kernel overflow
fix block shape error when nrows too big
2. Im2Col:
Support Batch in cuda
Support f32 to f32 both in cpu && cuda
3. DepthWiseConv:
Support by Im2Col && MulMat
4. Pool_2d:
Supoort avg pooling in cuda
5. HardSigmoid:
Imp in cuda
6. HardSwish:
Imp in cuda
* fix tabs instead of spaces
* code clean
* CUDA POOL2D
* ADD POOL2D test case in test-backend-ops.cpp
* code clean
* fix pool2d_kernel
nits
* fix bug in pool2d kernel
* fix avg pooling, count_include_pad
nits
* test-backend-ops : add more pool_2d tests
* cuda : fix warnings and formatting
* ggml : check types in release builds too in pool_2d
* test-backend-ops : remove f16 pool_2d tests
* cuda : more style fixes
* Add assert in ggml_cuda_op_pool2d
* pool2d float padding fallback
* test-backend-ops : add dst_type to im2col
---------
Co-authored-by: slaren <slarengh@gmail.com>
* Fix Vulkan F16 models
* Fix Vulkan context shift crash
* Add Vulkan to common.cpp dump_non_result_info_yaml function
* Fix bug in Vulkan CPY op
* Fix small matrix multiplication errors in AMD GPUs on Windows or with amdvlk
Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>
---------
Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>
* support SYCL backend windows build
* add windows build in CI
* add for win build CI
* correct install oneMKL
* fix install issue
* fix ci
* fix install cmd
* fix install cmd
* fix install cmd
* fix install cmd
* fix install cmd
* fix win build
* fix win build
* fix win build
* restore other CI part
* restore as base
* rm no new line
* fix no new line issue, add -j
* fix grammer issue
* allow to trigger manually, fix format issue
* fix format
* add newline
* fix format
* fix format
* fix format issuse
---------
Co-authored-by: Abhilash Majumder <30946547+abhilash1910@users.noreply.github.com>
* server : fix context shift + simplify self-extend
* server : take system_tokens into account
* server : more n_past fixes
* server : rever n_past_se changes
* Changed ugly xxd to literals.
HPP files are much more readable as multiline literals rather than hex arrays.
* Dashes in literal variable names.
Replace . and - with _ in file names -> variable names.
* Comment on removing xxd.
XXD-> string literals
* XXD to string literals.
Replaced these unreadable headers with string literal versions using new deps.sh.
* added cuda float16->float32 upcasting to ggml_cuda_cpy
* added ability to copy 4d tensors with the cuda backend
* added tests for float16_>float32 upcast and 4d tensor cuda copys
* added 4d copy test for float32->float16 copy
* applied patch suggested by @iamlemec
* simplify cpy tests
---------
Co-authored-by: slaren <slarengh@gmail.com>
* iq3_xxs: quantize/dequantize
RMSE seems a bit high-ish at about half-way between q2_K and
q3_K, so need to check more.
* iq3_xxs: CUDA dequantize works
* iq2_xxs: tuning quantization
* iq3_xxs: starting to look better
PPL on wiki.test.raw
LLaMA-v1-7B: 6.4218
LLaMA-v2-7B: 6.3560
Mistral-7B : 6.0717
This is better than Q3_K_XS, with a 5% reduction in quantized model
size.
* iq3_xxs: CUDA dot product
We have
PP-512: 5891 t/s
TG-128: 143.9 t/s
* iq3_xxs: scalar and AVX2 dot products
* iq3_xxs: ARM_NEON and Metal
Metal performance is decent, ARM_NEON is pathetic
* iq3_xxs: slightly better grid points
* Faster iq3_xxs and iq2_xs dot products on CUDA
* iq3_xxs: add some quant mix
* iq3_xxs: fix failing quantization test
Dot product still fails. Is this real?
* iq3_xxs: hopefully fix ROCm
* iq3_xxs: failing tests
This time the dot product accuracy did find an actual bug
in the AVX2 implementation.
* Add IQ3_XXS to test-backend-ops
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* allow empty --prompt-cache file
This allows the use of std::tmpnam(), std::tmpfile(), Python's tempfile.NamedTemporaryFile(), and similar create-empty-file API's for the user.
I switched from the C fopen API to the C++ filesystem api to get around the fact that, to the best of my knowledge, C has no portable way to get the file size above LONG_MAX, with std::ftell() returning long? fallback to std::ifstream for c++ < 17
(the project is currently targeting C++11 it seems - file_exists() and file_size() can be removed when we upgrade to c++17)
* formatting
(requested in codereview)
* remove c++17, file_is_empty
* Fix for a null pointer dereference if a metal GGML buffer fails to be allocated
* Freeing the allocated buffers rather than the pointer in ggml-alloc.c
* Fixed the fix of the fix
* Vulkan loader code
* Fix matmul kernel, continue implementation
* Continue implementation
* Vulkan memory management
* Vulkan development
* Matmul call
* Add aligned malloc and free for VMA
* Continue implementation
* First matmul success
* GEMM Kernel optimization
* 1D Blocktiling
* 2D Blocktiling
* Write coalescing
* Continue vulkan implementation and optimization
* First FP16 attempt, disabled for now
* Code abstraction, FP16 implementation, fix kernel, add FP16 to FP32 kernel
* Enable device extensions properly, restore fp16 matmul op
* Fix mulmat_f16
* Output FP32 in fp16 matmul shader
* Fix f16_to_f32 kernel
* dequant_q4_0 kernel
* Add VMA library
* Avoid requesting dedicated memory, VMA can decide that by itself
* Add bounds checking to matmul kernels, improve implementation, fix command buffers not freed properly
* add cmake commands
* Add 2d write operation, profiling code
* Fix 2d write
* Fix queue selection for AMD RADV
* Fix trailing whitespace in vk_mem_alloc.h
* Add WIP warp tile mat mul shaders
* Disable glslc optimization
* Disable glslc optimization for CMake
* Optimize warptile matmul shader, replace blocktile with it
* Add split-k optimization for small matrix multiplication
Use semaphores for synchronization instead of fences or waitidle
Rework async write/read for synchronization
* Fix validation errors, improve compatibility with AMD GPUs
* Rework command buffer handling
* Variable matmul kernel using specialization constants
* Fix synchronization on AMD, add barriers for buffer ownership transfer, add debug flag and prints
* Reuse semaphores
* Handle stage flags during command buffer submission properly
* Increase matmul test runs for consistent results
* Fix F32 matmul
* Add vectorized loading and zeropadding for matrix multiplication
* Use pinned memory for f16 preprocessing
* Don't force aligned matmul
* Don't free before queue done
* Replace VMA library with native Vulkan buffer management
* Basic offloading support with mul_f32 and dmmv for q4_0
* Run glslc commands in parallel
* Unroll loops in dmmv shader
* Reduce usage of waitIdle
* Reuse pinned allocation for f16 conversion
* Handle devices with only a single queue
* Fix trailing whitespace in CMakeLists.txt
* Allow parallel execution of kernels, parallelize third and fourth dimension calls
* Add fallback for devices only supporting one DescriptorSet per DescriptorPool
* Move to graph function similar to CUDA implementation
* Use F16 kernel for most things, replace q_f32 with mul_mat_q_f16 function
* Add F32 dmmv shaders
* Batch submissions
* Add .spv to gitignore
* Split off matrix vector multiplication for separate optimization
* Use single command buffer for matrix vector multiplication ops
* Reduce overhead of mul_f32 calls by using a single command buffer
* Add submission batching to mul_f32
* Fix tests
* Add missing barrier
* Add further missing barrier
* Add further ops
* Replace vk::QueueFamilyIgnored with VK_QUEUE_FAMILY_IGNORED to support more Vulkan header versions
* Remove unnecessary cblas link
* Fix descriptor set pre-allocation assert
* Add runtime shader compilation, start transferring shaders to this approach
* Transfer remaining shaders to header and compile on runtime
* Fix fp32 fallback if device doesn't support fp16, add force disable env var GGML_VULKAN_DISABLE_F16
* Add support for q4_1, q5_0, q5_1 and q8_0
* Remove unnecessary scalar layout extension
* Parse graph early to pre-record command buffers
* Add q6_k support
* Add multi-submit for command buffers
* Fix q6_k dequant shader for AMD
* Fix q6_k for GPUs without fp16 support
* Simplify q6_k fp16 fix
* Minor fixes
* Fix wg_denom of m-mulmat shaders
* Add Python-based Vulkan shader generator
* Replace shaderc dependency with precompiled shaders
Fix python script to generate shaders
* Clean up code
* Fix shader generator script Windows compatibility
Co-authored-by: Concedo <39025047+LostRuins@users.noreply.github.com>
* Close file before deletion
* Fix vulkan shader fp32 name
* Add q2_k and q3_k support
Add validation check to compare shader results to cpu results
* Add q4_k support
* Add q5_k support
* Bake SPIR-V bytecode into the library instead of loading shaders from file
* Switch to signal semaphores for flexibility
Prepare broadcasting support for mul mat
* Finish broadcasting mul mat support for GQA
* Clean up unused functions
Add repeat op
* Add further ops, not yet enabled. Improve semaphore code
* Reduce number of used semaphores by utilizing timelines more properly
* Remove queue information
* Reuse timeline semaphores, allow parallel operation with binary semaphores to work around nvidia driver limitations
* Add Vulkan to llama-bench
* Remove cblas dependency
* Fix matmul k-split bug
* Fix q4_k dmmv K_QUANTS_PER_ITERATION 1 shader
* Add RMS Norm shader, rework op_f32 shader setup, fix matmul bug
* Fix issues with float16 overflows in shaders
* Fix issues with older Vulkan headers on Ubuntu 22.04
* Allow multi-op partial offloading by parsing the graph to preallocate enough between-op buffers
* Implement further ops, rework op_f32 calls, fix bugs
* Finish full offloading support, add last remaining ops, fix bugs, remove redundant code
* Upload generated file ggml-vulkan-shaders.hpp, remove redundant shaders
* Merge upstream changes, fix conflicts, adapt soft_max op
* Fix Python and shader header format
* Free model gpu buffers on exit
* Use single queue per device to simplify code
* Add matmul shader support for running multiple calculations in parallel
* Switch from semaphore-synchronized multiple command buffers per op to single command buffer for multiple ops, whole graph if possible
* Fix missing event cast
* Replace uint64_t(-1) with UINT64_MAX, rename function for clarity
* Fix warning about empty C function parameters
* Fix compiler warnings
* Properly implement Vulkan backend buffer handling
* Fix oversized host staging buffers
* Simplify barrier synchronization calls
* Fix gcc warnings
* Implement max_size for backend buffer types to limit the size of a single allocation
* Use min of maxMemoryAllocationSize and maxBufferSize for device max allocation size
* refactor multi buf
* Disable unsupported ops to fix tests
* Check for maintenance4 support before using it
* Handle devices with only a single queue
* Fix single queue logic
* propagate buffer usage in multi buffers
* Implement rope_neox op
* Cleanup header and other files
* Simplify gpu_extras by removing events and putting staging memcpys into contexts
* Move queue into context
Add not-yet-enabled async backend ops
* Simplify context use, optimize matmul shader for warp size 64 (AMD GCN), fix split_k matmul shader optimization
* Add get_max_size to SYCL backend.
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama : fix trailing whitespace
---------
Co-authored-by: Henri Vasserman <henv@hot.ee>
Co-authored-by: Concedo <39025047+LostRuins@users.noreply.github.com>
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* first update for migration
* update init_cublas
* add debug functio, commit all help code
* step 1
* step 2
* step3 add fp16, slower 31->28
* add GGML_LIST_DEVICE function
* step 5 format device and print
* step6, enhance error check, remove CUDA macro, enhance device id to fix none-zero id issue
* support main device is non-zero
* step7 add debug for code path, rm log
* step 8, rename all macro & func from cuda by sycl
* fix error of select non-zero device, format device list
* ren ggml-sycl.hpp -> ggml-sycl.h
* clear CMAKE to rm unused lib and options
* correct queue: rm dtct:get_queue
* add print tensor function to debug
* fix error: wrong result in 658746bb26702e50f2c59c0e4ada8e9da6010481
* summary dpct definition in one header file to replace folder:dpct
* refactor device log
* mv dpct definition from folder dpct to ggml-sycl.h
* update readme, refactor build script
* fix build with sycl
* set nthread=1 when sycl, increase performance
* add run script, comment debug code
* add ls-sycl-device tool
* add ls-sycl-device, rm unused files
* rm rear space
* dos2unix
* Update README_sycl.md
* fix return type
* remove sycl version from include path
* restore rm code to fix hang issue
* add syc and link for sycl readme
* rm original sycl code before refactor
* fix code err
* add know issue for pvc hang issue
* enable SYCL_F16 support
* align pr4766
* check for sycl blas, better performance
* cleanup 1
* remove extra endif
* add build&run script, clean CMakefile, update guide by review comments
* rename macro to intel hardware
* editor config format
* format fixes
* format fixes
* editor format fix
* Remove unused headers
* skip build sycl tool for other code path
* replace tab by space
* fix blas matmul function
* fix mac build
* restore hip dependency
* fix conflict
* ren as review comments
* mv internal function to .cpp file
* export funciton print_sycl_devices(), mv class dpct definition to source file
* update CI/action for sycl code, fix CI error of repeat/dup
* fix action ID format issue
* rm unused strategy
* enable llama_f16 in ci
* fix conflict
* fix build break on MacOS, due to CI of MacOS depend on external ggml, instead of internal ggml
* fix ci cases for unsupported data type
* revert unrelated changed in cuda cmake
remove useless nommq
fix typo of GGML_USE_CLBLAS_SYCL
* revert hip cmake changes
* fix indent
* add prefix in func name
* revert no mmq
* rm cpu blas duplicate
* fix no_new_line
* fix src1->type==F16 bug.
* pass batch offset for F16 src1
* fix batch error
* fix wrong code
* revert sycl checking in test-sampling
* pass void as arguments of ggml_backend_sycl_print_sycl_devices
* remove extra blank line in test-sampling
* revert setting n_threads in sycl
* implement std::isinf for icpx with fast math.
* Update ci/run.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/sycl/run-llama2.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/sycl/run-llama2.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* add copyright and MIT license declare
* update the cmd example
---------
Co-authored-by: jianyuzh <jianyu.zhang@intel.com>
Co-authored-by: luoyu-intel <yu.luo@intel.com>
Co-authored-by: Meng, Hengyu <hengyu.meng@intel.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* feat: add Dockerfiles for each platform that user ./server instead of ./main
* feat: update .github/workflows/docker.yml to build server-first docker containers
* doc: add information about running the server with Docker to README.md
* doc: add information about running with docker to the server README
* doc: update n-gpu-layers to show correct GPU usage
* fix(doc): update container tag from `server` to `server-cuda` for README example on running server container with CUDA
* Support for Yi-VL, templating fix for mobileVLM
* ws
* Update examples/llava/clip.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update llava-cli.cpp
* Update clip.cpp
bugfix for new conversions
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
updated the n_task calculation to use max number of
threads possible. This has improved the prompt eval
performance by around 5% for DOT kernels and by
around 10% for MMLA kernels on AWS Graviton3.
* server: add llama_server_queue struct
* server: add llama_server_response_event
* server: add comments
* server: move all mutexes away from server.cpp
* server: correct multitask response
* server: only add back deferred tasks when one slot is available
* server: fix a race condition cause by "request_completion"
* scripts : add lib.sh and lib_test.sh
* scripts : stub out new ci-run.sh script
* scripts : switch to PascalCase for functions
This looks a little odd at first, but I find it very useful as a
convention to know if a command is part of our code vs a builtin.
* scripts : add some fancy conversion from snake_case to PascalCase
* Add venv to ci/run.sh
* Revert scripts work
* scripts : add wrapper script for local use of ci/run.sh
* Simplify .gitignore for tests, clang-tidy fixes
* Label all ctest tests
* ci : ctest uses -L main
* Attempt at writing ctest_with_model
* Update test-model-load-cancel
* ci : add ctest_with_model for debug and release
ggml-ci
* Fix gg_get_model function
ggml-ci
* got stuck on CMake
* Add get_model.cpp to tests/CMakeLists.txt
ggml-ci
* Fix README.md output for ctest_with_model
ggml-ci
* workflows : use `-L main` for all ctest
ggml-ci
* Fixes
* GG_RUN_CTEST_MODELFILE => LLAMACPP_TESTMODELFILE
* Always show warning rather than failing if model file variable is not
set
* scripts : update usage text for ci-run.sh
* implemented dynamic temperature sampling from koboldcpp
* removed trailing whitespace
* removed unused temp parameter in llama_sample_entropy
* exposed exponent_val in dynamic temp sampler
* added debug check for printf statements
* use nullptr in llama_sample_softmax call during llama_sample_entropy
this avoids counting the time taken stats twice
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* return earlier if there is only 1 candiate (i.e. max_entropy == 0)
* reformat 't' case in llama_sample_queue
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* check for one or zero candidates case in llama_sample_entropy
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
this fixes the error I encountered when trying to run the convert.py
script in a venv:
```
$ nix develop
[...]$ source .venv/bin/activate
(.venv)
[...]$ pip3 install -r requirements.txt
<... clipped ...>
[...]$ python3 ./convert.py
Traceback (most recent call last):
File "/home/mhueschen/projects-reference/llama.cpp/./convert.py", line 40, in <module>
from sentencepiece import SentencePieceProcessor
File "/home/mhueschen/projects-reference/llama.cpp/.venv/lib/python3.11/site-packages/sentencepiece/__init__.py", line 13, in <module>
from . import _sentencepiece
ImportError: libstdc++.so.6: cannot open shared object file: No such file or directory
```
however, I am not sure this is the cleanest way to address this linker
issue...
* kl-divergence: be able to save all logits to a file
* Add ability to compute KL-divergence
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* make GGML_TASK_INIT phase can be run in multithread
* multithreaded dequantize in mul_mat when using blas library
* minor fixes
* update outdated comment
* fix coding style
* simplify code
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* MobileVLM native implementation
* delete depthwise_conv_2d and permute_cpy relative code, replace the two by the existed functions, and opt ldp definition, support LLAMA_PERF option for CMake
* move android script to example/llava directory
* Fix the editor config checks
---------
Co-authored-by: Chenxiaotao03 <chenxiaotao03@meituan.com>
* llama : support StableLM 2 1.6B
* convert : fix Qwen's set_vocab wrongly naming all special tokens [PAD{id}]
* convert : refactor Qwen's set_vocab to use it for StableLM 2 too
* nix : add tiktoken to llama-python-extra
* convert : use presence of tokenizer.json to determine StableLM tokenizer loader
It's a less arbitrary heuristic than the vocab size.
This commit adds `--sample-start` and `--include-sample-start` to the
output from the main function in finetune.cpp.
The motivation for this is that even though these are set explicitly by
the user via the command line, if one forgets to set them then it is
useful to have their values printed out. Otherwise it is possible to go
through the whole training process before realizing that the values are
not what one expected.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Add Q3_K_XS - intermediate size between Q2_K and Q3_K_S
* Q3_K_XS: quanize first 1/8 of ffn_down layers with Q4_K
Together with an importance matrix, this brings perplexity
for LLaMA-v2-70B below the perplexity of the former Q2_K
with a 800 MB smaller quantized model size.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* TruthfulQA: 1st attempt, does not look like it is working
The same implementation can be used for HellaSwag as well,
so I converted a HellaSwag validation dataset to the binary
format used here and tested with that. The score is only
around 50, so something is not quite right.
* TruthfulQA: works but the result is bad
I know it works because if I convert the HellaSwag validation
data to the binary format used in the truthful_qa_score() function
I get the exact same result as from the hellaswag_score() function.
But I guess, the questions are tricky and the way I have done
the combination of question + answer is very likely not the best.
The TruthfulQA validation dataset contains 817 questions, with
random chance result around 19%. With this version I get
29.1% for Mistral-7B and 55.2% for Mistral-7B-Instruct-v0.2.
The HF leader board results for these two models are
42.2% and 68.3%, respectively.
* TruthfulQA: fix random sample
* TruthfulQA: prepare tasks in parallel for large test datasets
* Rename truthful_qa to multiple_choice
* Make MSVC happy
I had forgotten that MSVC does not make constexpr's available
inside a lambda.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
PR #4818 (merged last week) reintroduced a config check for vocab_size that was addressed in PR #4258 (merged 2023-11-30).
Without the fix, llama2 models can't be converted. The error is:
`ValueError: The model's vocab size is set to -1 in params.json. Please update it manually. Maybe 32000?`
For Mistral-7B and fp16, time on my system goes down from 536 seconds
to 423 seconds for the full evaluation dataset (10042 tasks).
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* winogrande: simple implementation
It doesn't look like it is working - why?
For Mistral-7B it is barely better than
random chance (score ~60% for 1267 tasks), while I see
Mistral-7B scoring 78.4% on the HF leader board.
1-sigma statistical uncertainty for 1267 tasks is ~1.4,
so no way the difference is due to statistics.
* winogrande: somewhat better
Score for Mistrali7-B is now 68.9 on the validation set of
winogrande_debiased. Still far from the reported 78.4, but
better than what I had before.
* winogrande: improving
Mistral-7B score is now 73.56.
Still not quite 78.4 but getting there.
We are also getting a lower score on HellaSwag
compared to HF leader board, so I'm not expecting
we will get up to 78.4 anyway.
It looks like it is better to skip the choice word(s)
when evaluating the average log-likelihood. This kind of
makes sense because a more common word (in Winogrande this is
often a name) will have a higher probability without knowing
about the follow up context, and this will skew the log-likelihood
towards the more common word. We can only do this if the
choice words are not last in the sentence.
It also looks like it is better to skip the punctuation at the
end of the sentence, provided the choice words are not last.
* winogrande: add dataset instructions
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Metal memory: Small memory leak on init, dangling pointer, and unused autorelease pool in graph compute
* SPM header potential fix
* Reverting symlinks
* ggml : add IQ2 to test-backend-ops + refactoring
ggml-ci
* cuda : update supports_op for IQ2
ggml-ci
* ci : enable LLAMA_CUBLAS=1 for CUDA nodes
ggml-ci
* cuda : fix out-of-bounds-access in `mul_mat_vec_q`
ggml-ci
* tests : avoid creating RNGs for each Q tensor
ggml-ci
* tests : avoid creating RNGs for each tensor
ggml-ci
* backend : add eval callback
ggml-ci
* backend : group nodes in a single compute when user don't need them
* backend : clean-up the implementation
ggml-ci
* simple : do not perform tensor data copy if not needed
* simple : fix
* imatrix : offload to GPU support
* imatrix : fix ggml_mul_mat_id hanlding
ggml-ci
* ci : add imatrix test
ggml-ci
* ci : rearrange output
ggml-ci
* backend : add eval callback
ggml-ci
* backend : group nodes in a single compute when user don't need them
* backend : clean-up the implementation
ggml-ci
* simple : do not perform tensor data copy if not needed
* simple : fix
* simple : no need for ggml_is_contiguous + fix bool parse
* llama : fix callback placement in llama_context_params
* backend : avoid double-ask callback calls
* simple : restore examples, imatrix will serve as a demo
This commit adds the name of the training data file to the log message
printed when the training data is tokenized.
The motivation for this change is that it can be useful to show which
file is being tokenized when running the finetune example.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Introduce starter project for Android
Based on examples/llama.swiftui.
* Add github workflow
* Set NDK version
* Only build arm64-v8a in CI
* Sync bench code
* Rename CI prop to skip-armeabi-v7a
* Remove unused tests
* metal: Log `recommendedMaxWorkingSetSize` on iOS 16+
* Only log on iOS and macOS, ignoring tvOS and other platforms
* Check for Xcode version before using recommendedMaxWorkingSetSize
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This change makes it possible to build ggml-cuda.cu and ggml-metal.m as
independent dynamic shared objects, that may be conditionally linked at
runtime in a multiplatform binary. It introduces a GGML_CALL annotation
that documents which functions have a cyclic call relationship, between
the application code and GPU modules.
This change does nothing, unless the build defines -DGGML_MULTIPLATFORM
which causes back-references and function pointers to conform to MS ABI
which is supported by NVCC, ROCm, XCode, GCC and Clang across platforms
This commit replaces the magic number LLAMA_FILE_MAGIC_LORA used in
finetune.cpp with LLAMA_FILE_MAGIC_GGLA defined in llama.h.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Fix ffn_down quantization mix for MoE models
In #4872 I did not consider the part where every third
tensor is quantized with more bits. Fir MoE this leads to tensors
of the same layer being quantized with different number of bits,
which is not considered as a possibility in the inference implementation
(it is assumed all experts use the same quantization).
* Fix the fix
* Review suggestion
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* examples : save-load-state: save only required state
* llama : only reserve n_vocab * n_batch at most for logits
llama_decode asserts that only n_batch tokens are passed each call, and
n_ctx is expected to be bigger than n_batch.
* llama : always reserve n_vocab * n_batch for logits
llama_context de-serialization breaks if the contexts have differing
capacity for logits and llama_decode will at maximum resize to
n_vocab * n_batch.
* llama : only save and restore used logits
for batch sizes of 512 this reduces save state in the best case by
around 62 MB, which can be a lot if planning to save on each message
to allow regenerating messages.
* llama : use ostringstream and istringstream for save and load
* llama : serialize rng into minimum amount of space required
* llama : break session version due to serialization changes
* add the parameter : --no-display-prompt , combine with --log-disable it will display only the generated tokens
* remove empty line
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* metal : detect more GPU families
* metal : refactor kernel loading
* metal : set kernel family requirements
* metal : fix kernel init + fix compile options
* metal : take into account simdgroup reduction support
* metal : print only skipped kernels
* metal : fix check for simdgroup reduction support
* metal : check for Metal 3
* metal : free allocations
* metal : normalize encoder:setComputePipelineStatus calls
ggml-ci
* metal : fix Metal3 family check
ggml-ci
* metal : check for simdgroup matrix mul. feature
ggml-ci
* llama : ggml-backend integration
* ggml-backend : add names to buffers
* fix unmap after loading
* batched-bench : add tensor_split param
* llama : check for null tensor_split
* ggml-backend : increase GGML_MAX_BACKENDS
* improve graph splitting, partial fix for --no-kv-offload
* cuda : add ggml-backend split buffer support
* cuda : do not create buffer types for devices that don't exist (fixes usage without CUDA devices available)
* ggml : fix null backend dereference (#4807)
* ggml : fix null backend dereference
* ggml : also check ggml_backend_is_cpu
* test-backend-ops : check buffer allocation failures
* llama : add cparam (split_mode) and command line argument (--split-mode, -sm) to configure the split mode (none, layer or row)
* ggml : fix mul_mat_id work size
* llama : rewrite session kv load/set without graphs
* minor
* llama : only initialize used backends, free backends on context free
* llama : abort ctx if cuda backend init fails
* llama : rewrite lora with ggml-backend and compute on CPU
ggml-ci
* llama : only map to a backend buffer the region of the file mapping containing the tensors used in the buffer
* opencl : add ggml-backend buffer type
* cuda : only use batched_cublas with batched mat muls (fixes fp16 tg perf)
* llama : on Metal, by default offload the full model
ggml-ci
* metal : page align the data ptr (#4854)
* Apply suggestions from code review
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* cuda : fix split buffer free
* address review comments
* llama-bench : add split-mode parameter
* fix whitespace
* opencl : fix double initialization
* server : add --split-mode parameter
* use async copy and compute to improve multi-gpu performance
ggml-ci
* use async memcpys to copy the graph outputs to the CPU
* fix opencl
* use a host buffer for the cpu compute buffer for faster copies to the gpu
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
This commit replaces the magic number used in export-lora.cpp with
the one defined in llama.h, which is indirectly included via common.h.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Updated Models Layout
- Added a models drawer
- Added downloading directly from Hugging Face
- Load custom models from local folder
- Delete models by swiping left
* trimmed trailing white space
* Updated Models Layout
* common : streamline the formatting of help
- Separate alternative parameters by a comma
- Do not indent `--version` differently
* Update common/common.cpp
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Restore intended k-quants quantization mixes for MoE models
* Update Q2_K_S values in the quantize tool
Still using LLaMA-v1 PPL values in the quant description
today does not make much sense. But let's leave this update
for another PR.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* iq2_xs: basics
* iq2_xs: this should have been in the basics
* iq2_xs: CUDA and scalar CPU works
* iq2_xs: WIP Metal
* iq2_xs: Metal now works
* iq2_xs: working, but dog slow, ARM_NEON dot product
* iq2_xs: better ARM_NEON dot product
We are now at 19.5 t/s for TG-128 and 61 t/s for PP-512 when
running on the CPU.
* iq2_xs: AVX2 dot product - 19.5 t/s
* iq2_xs: faster AVX2 dit product
21.4 t/s for TG-128, 59.2 t/s for PP-512.
The latter is 2x compared to the previous version.
* iq2_xs: had forgotten to delete iq2-data.h
* Add llama enum for IQ2_XS
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* server: added support for multiple api keys, added loading api keys from file
* minor: fix whitespace
* added file error handling to --api-key-file, changed code to better
reflect current style
* server: update README.md for --api-key-file
---------
Co-authored-by: Michael Coppola <info@michaeljcoppola.com>
* added /health endpoint to the server
* added comments on the additional /health endpoint
* Better handling of server state
When the model is being loaded, the server state is `LOADING_MODEL`. If model-loading fails, the server state becomes `ERROR`, otherwise it becomes `READY`. The `/health` endpoint provides more granular messages now according to the server_state value.
* initialized server_state
* fixed a typo
* starting http server before initializing the model
* Update server.cpp
* Update server.cpp
* fixes
* fixes
* fixes
* made ServerState atomic and turned two-line spaces into one-line
* updated `server` readme to document the `/health` endpoint too
* used LOG_INFO after successful model loading
NULL can be an integer constant expression with the value zero, in this case the behavior would be undefined because of an incorrect type being passed to the variable arguments.
* added /health endpoint to the server
* added comments on the additional /health endpoint
* Better handling of server state
When the model is being loaded, the server state is `LOADING_MODEL`. If model-loading fails, the server state becomes `ERROR`, otherwise it becomes `READY`. The `/health` endpoint provides more granular messages now according to the server_state value.
* initialized server_state
* fixed a typo
* starting http server before initializing the model
* Update server.cpp
* Update server.cpp
* fixes
* fixes
* fixes
* made ServerState atomic and turned two-line spaces into one-line
* updated `server` readme to document the `/health` endpoint too
* added /health endpoint to the server
* added comments on the additional /health endpoint
* Better handling of server state
When the model is being loaded, the server state is `LOADING_MODEL`. If model-loading fails, the server state becomes `ERROR`, otherwise it becomes `READY`. The `/health` endpoint provides more granular messages now according to the server_state value.
* initialized server_state
* fixed a typo
* starting http server before initializing the model
* Update server.cpp
* Update server.cpp
* fixes
* fixes
* fixes
* made ServerState atomic and turned two-line spaces into one-line
This update categorizes models with 24 layers as MODEL_1B, ensuring compatibility with different Phi model variants without impacting existing Phi-2 model functionality.
Uses ggml functions instead of hardcoded names and adds support to quantize into the modern Q-K variants.
This is just the bare minimum to get k-types working - a more refined choice of types would be needed to get best quality on low quantizations.
I ran a few tests, it doesn't break anything I could notice and a Q6_K ViT works almost as well as Q8_0 but 3 times the inference speed.
* Update Imports and Add Notes for Future Reference
- Updated import statements in `convert.py`.
- Added import for `AutoTokenizer` from `transformers` module.
- Added conditional import for `gguf` from the local directory.
- Added comments and notes for future reference.
Additional Notes:
- Noted removal of a redundant `TypeAlias` import.
- Noted the removal of a `gguf` debug statement.
- Commented on the presence of `ARCH` and `NDArray` definitions.
- Commented on cleaning up and refactoring data type definitions.
* Refine Model Hyperparameters and Params Class
- Updated type annotations to use `Optional` for clarity.
- Improved method names and attribute consistency.
- Removed unnecessary variables for better code readability.
Additional Notes:
- Highlighted the use of `Optional` for clearer intent.
- Ensured backward and forward compatibility.
* Restore BpeVocab and SentencePieceVocab classes
- Restored the BpeVocab class for handling BPE tokenization.
- Restored the SentencePieceVocab class for SentencePiece tokenization.
These classes are essential for maintaining the original behavior of the codebase.
* refactor: Standardize vocabulary handling with HfVocab
- Replaced VocabLoader with HfVocab, aligning vocabulary handling across classes.
- Updated initialization of HfVocab with local_files_only=True for AutoTokenizer.
- Introduced optional parameter fname_added_tokens for flexible added token management.
- Streamlined added token handling for clarity and conciseness.
- Maintained special tokens and IDs, enhancing token management.
- Simplified token processing methods for improved readability.
- Added a placeholder for score computation with a default value of -1000.0.
- Optimized newline token check for efficiency.
- Updated __repr__ function for clarity in representation.
- Adjusted type alias Vocab to include BpeVocab, SentencePieceVocab, and HfVocab.
- Removed redundant code related to special token handling, reverse vocabulary mapping, and vocabulary file detection.
This refactoring promotes a standardized and modular approach to vocabulary management, facilitating future integration with a VocabFactory and improving code maintainability and scalability.
* refactor: Enhance readability, functionality, and code quality
- Improved code formatting and readability for better maintainability.
- Refactored LazyUnpickler's CLASSES dictionary for clarity.
- Added print statements and warnings in check_vocab_size for user feedback.
- Removed find_vocab_file_path, as it's superseded by VocabFactory.
- Preparatory changes for upcoming classes: OutputFile and VocabFactory.
- Overall focus on code quality, error handling, and consistency.
These changes reflect a continuous effort to refine the codebase, ensuring it meets best practices and prepares for future enhancements, such as the VocabFactory.
* refactor: Update OutputFile class for enhanced model vocabulary management
- Restructured the constructor for improved readability.
- Updated `add_meta_arch` method for flexible model name determination.
- Introduced `handle_tokenizer_model` for mapping vocab types to supported tokenizer models.
- Streamlined vocabulary extraction with `extract_vocabulary_from_model`.
- Simplified vocabulary metadata addition using `add_meta_vocab`.
- Refactored `add_tensor_info` for clarity and consistency.
- Improved error handling for better user feedback.
These changes signify the development of a versatile and comprehensive `OutputFile` class, enabling efficient management of model conversion output, metadata, vocabulary, and tensor information.
* feat: Introduce VocabFactory for flexible vocabulary management in model conversion
- The VocabFactory class is added to facilitate modular vocabulary handling.
- The constructor initializes a directory path and detects vocabulary-related files.
- The _select_file method provides file paths based on vocabulary type (e.g., BPE, SentencePiece).
- _create_special_vocab generates special vocabularies, accommodating different types.
- The load_vocab method loads vocabularies, handling BPE, SentencePiece, and Hugging Face Fast Tokenizer.
- Error handling and logging enhance debugging and user feedback.
- The modular and flexible design simplifies vocabulary management and supports future extensions.
The VocabFactory class enhances code modularity and maintainability, allowing versatile vocabulary handling in the model conversion process.
* refactor: Improve code organization, argument parsing, and user interface
- Renamed 'default_outfile' to 'default_output_file' for clarity.
- Refactored argument parser setup into 'get_argument_parser' function.
- Introduced descriptive comments for each argument in the parser.
- Added '--vocab-type' argument with choices ["spm", "bpe", "hfft"] for vocabulary processing.
- Improved flag naming consistency: '--outfile' to '--out-file' and '--bigendian' to '--big-endian'.
- Enhanced error handling to prevent overwriting input data in 'default_output_file'.
- Made 'argv' in 'main' an optional parameter for flexibility.
- Introduced dynamic import for 'awq.apply_awq' based on 'args.awq_path' for conditional dependency.
These changes enhance code clarity, organization, and the user interface of the script, aligning it with Python best practices and improving maintainability.
* refactor: Further refine functionality, improve user interaction, and streamline vocabulary handling
- Renamed command-line arguments for clarity and consistency.
- Improved path resolution and import adjustments for robustness.
- Thoughtfully handled 'awq-path' and conditional logic for the weighted model.
- Enhanced model and vocabulary loading with the 'VocabFactory' class for structured and adaptable loading.
- Strengthened error handling and user feedback for a more user-friendly experience.
- Structured output file handling with clear conditions and defaults.
- Streamlined and organized the 'main' function for better logic flow.
- Passed 'sys.argv[1:]' to 'main' for adaptability and testability.
These changes solidify the script's functionality, making it more robust, user-friendly, and adaptable. The use of the 'VocabFactory' class is a notable enhancement in efficient vocabulary handling, reflecting a thoughtful and iterative approach to script development.
* chore: Apply ruff formatting to convert.py
Signed-off-by: teleprint-me <77757836+teleprint-me@users.noreply.github.com>
* Revert to commit 0614c33
* chore: Apply flake8 formatting rules
Signed-off-by: teleprint-me <77757836+teleprint-me@users.noreply.github.com>
* refactor: Revise `check_vocab_size` for Enhanced Clarity and Correctness
- Resolved an unreachable branch issue by reorganizing the conditional structure.
- Moved the special case check for `params.n_vocab == -1` to the top for immediate assertion.
- Flattened the conditional logic for improved clarity and predictability of the function's behavior.
These changes enhance the readability and functional correctness of the `check_vocab_size` function without altering its intended functionality.
* py : fix outfile and outtype
* py : suggest hint for missing vocab size
---------
Signed-off-by: teleprint-me <77757836+teleprint-me@users.noreply.github.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This change fixes an issue where supplying `--image missing-file` would
result in a segfault due to a null pointer being dereferenced. This can
result in distracting info being printed if robust crash analysis tools
are being used.
* updated server readme to reflect the gg/server-token-probs-4088 commit
added explanation for the API's completion result which now includes `completion_probabilities`. Also added a JSON schema that shows the type/structure of `completion_probabilities`.
* simplified the `completion_probabilities` JSON schema
It's now easier to understand what the structure of `completion_probabilities` looks like.
* minor : fix trailing whitespace
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* iq2_xxs: basics
* iq2_xxs: scalar and AVX2 dot products
Needed to change Q8_K to have quants in the -127...127 range,
else the IQ2_XXS AVX implementation becomes very awkward.
The alternative would have been to use Q8_0 instead. Perhaps
I'll change later, for now this is what we have.
* iq2_xxs: ARM_NEON dot product
Somehow strangely slow (112 ms/token).
* iq2_xxs: WIP Metal
Dequantize works, something is still wrong with the
dot product.
* iq2_xxs: Metal dot product now works
We have
PP-512 = 475 t/s
TG-128 = 47.3 t/s
Not the greatest performance, but not complete garbage either.
* iq2_xxs: slighty faster dot product
TG-128 is now 48.4 t/s
* iq2_xxs: slighty faster dot product
TG-128 is now 50.9 t/s
* iq2_xxs: even faster Metal dot product
TG-128 is now 54.1 t/s.
Strangely enough, putting the signs lookup table
into shared memory has a bigger impact than the
grid values being in shared memory.
* iq2_xxs: dequantize CUDA kernel - fix conflict with master
* iq2_xxs: quantized CUDA dot product (MMVQ)
We get TG-128 = 153.1 t/s
* iq2_xxs: slightly faster CUDA dot product
TG-128 is now at 155.1 t/s.
* iq2_xxs: add to llama ftype enum
* iq2_xxs: fix MoE on Metal
* Fix missing MMQ ops when on hipBLAS
I had put the ggml_supports_mmq call at the wrong place.
* Fix bug in qequantize_row_iq2_xxs
The 0.25f factor was missing.
Great detective work by @ggerganov!
* Fixing tests
* PR suggestion
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* ggml : disable fast-math for Metal (cmake build only)
ggml-ci
* metal : fix Metal API debug warnings
* cmake : add -fno-inline for Metal build (#4545)
* metal : fix API debug warnings
* metal : fix compile warnings
* metal : use uint64_t for strides
* cmake : rename option to LLAMA_METAL_SHADER_DEBUG
* metal : fix mat-vec Q8_0 kernel for BS > 1
* metal : normalize mat-vec kernel signatures
* cmake : respect LLAMA_QKK_64 option
* metal : fix mat-vec Q4_K kernel for QK_K == 64
* metal : optimizing ggml_mul_mat_id (wip)
* metal : minor fix
* metal : opt mul_mm_id
* Add n_key_dim and n_value_dim
Some models use values that are not derived from `n_embd`.
Also remove `n_embd_head` and `n_embd_gqa` because it is not clear
which "head" is referred to (key or value).
Fix issue #4648.
* Fix `llm_build_kqv` to use `n_value_gqa`
* Rebase
* Rename variables
* Fix llm_build_kqv to be more generic wrt n_embd_head_k
* Update default values for n_embd_head_k and n_embd_head_v
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Fix llm_load_tensors: the asserts were not backcompat
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Changes to server to allow metadata override
* documentation
* flake.nix: expose full scope in legacyPackages
* flake.nix: rocm not yet supported on aarch64, so hide the output
* flake.nix: expose checks
* workflows: nix-ci: init; build flake outputs
* workflows: nix-ci: add a job for eval
* workflows: weekly `nix flake update`
* workflows: nix-flakestry: drop tag filters
...and add a job for flakehub.com
* workflows: nix-ci: add a qemu job for jetsons
* flake.nix: suggest the binary caches
* flake.lock: update
to a commit recently cached by nixpkgs-cuda-ci
---------
Co-authored-by: John <john@jLap.lan>
Co-authored-by: Someone Serge <sergei.kozlukov@aalto.fi>
* ggml : disable fast-math for Metal (cmake build only)
ggml-ci
* metal : fix Metal API debug warnings
* cmake : add -fno-inline for Metal build (#4545)
* metal : fix API debug warnings
* metal : fix compile warnings
* metal : use uint64_t for strides
* cmake : rename option to LLAMA_METAL_SHADER_DEBUG
* metal : fix mat-vec Q8_0 kernel for BS > 1
* metal : normalize mat-vec kernel signatures
* cmake : respect LLAMA_QKK_64 option
* metal : fix mat-vec Q4_K kernel for QK_K == 64
ggml-ci
* feat: add avx_vnni based on intel documents
* ggml: add avx vnni based on intel document
* llama: add avx vnni information display
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* Update ggml.c
Fix indentation upgate
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* python: add check-requirements.sh and GitHub workflow
This script and workflow forces package versions to remain compatible
across all convert*.py scripts, while allowing secondary convert scripts
to import dependencies not wanted in convert.py.
* Move requirements into ./requirements
* Fail on "==" being used for package requirements (but can be suppressed)
* Enforce "compatible release" syntax instead of ==
* Update workflow
* Add upper version bound for transformers and protobuf
* improve check-requirements.sh
* small syntax change
* don't remove venvs if nocleanup is passed
* See if this fixes docker workflow
* Move check-requirements.sh into ./scripts/
---------
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
* flake.lock: update to hotfix CUDA::cuda_driver
Required to support https://github.com/ggerganov/llama.cpp/pull/4606
* flake.nix: rewrite
1. Split into separate files per output.
2. Added overlays, so that this flake can be integrated into others.
The names in the overlay are `llama-cpp`, `llama-cpp-opencl`,
`llama-cpp-cuda`, and `llama-cpp-rocm` so that they fit into the
broader set of Nix packages from [nixpkgs](https://github.com/nixos/nixpkgs).
3. Use [callPackage](https://summer.nixos.org/blog/callpackage-a-tool-for-the-lazy/)
rather than `with pkgs;` so that there's dependency injection rather
than dependency lookup.
4. Add a description and meta information for each package.
The description includes a bit about what's trying to accelerate each one.
5. Use specific CUDA packages instead of cudatoolkit on the advice of SomeoneSerge.
6. Format with `serokell/nixfmt` for a consistent style.
7. Update `flake.lock` with the latest goods.
* flake.nix: use finalPackage instead of passing it manually
* nix: unclutter darwin support
* nix: pass most darwin frameworks unconditionally
...for simplicity
* *.nix: nixfmt
nix shell github:piegamesde/nixfmt/rfc101-style --command \
nixfmt flake.nix .devops/nix/*.nix
* flake.nix: add maintainers
* nix: move meta down to follow Nixpkgs style more closely
* nix: add missing meta attributes
nix: clarify the interpretation of meta.maintainers
nix: clarify the meaning of "broken" and "badPlatforms"
nix: passthru: expose the use* flags for inspection
E.g.:
```
❯ nix eval .#cuda.useCuda
true
```
* flake.nix: avoid re-evaluating nixpkgs too many times
* flake.nix: use flake-parts
* nix: migrate to pname+version
* flake.nix: overlay: expose both the namespace and the default attribute
* ci: add the (Nix) flakestry workflow
* nix: cmakeFlags: explicit OFF bools
* nix: cuda: reduce runtime closure
* nix: fewer rebuilds
* nix: respect config.cudaCapabilities
* nix: add the impure driver's location to the DT_RUNPATHs
* nix: clean sources more thoroughly
...this way outPaths change less frequently,
and so there are fewer rebuilds
* nix: explicit mpi support
* nix: explicit jetson support
* flake.nix: darwin: only expose the default
---------
Co-authored-by: Someone Serge <sergei.kozlukov@aalto.fi>
This change makes it possible to use flags like `--grammar` when using
the `llava-cli` program. The rest is just code cleanup deleting a long
standing TODO comment.
This change also ensures that logging information is emitted to stderr
which helps the `llava-cli` command be more friendly to shell scripts.
See Mozilla-Ocho/llamafile@1cd334f
The server currently schedules tasks using a sleep(5ms) busy loop. This
adds unnecessary latency since most sleep implementations do a round up
to the system scheduling quantum (usually 10ms). Other libc sleep impls
spin for smaller time intervals which results in the server's busy loop
consuming all available cpu. Having the explicit notify() / wait() code
also helps aid in the readability of the server code.
See mozilla-Ocho/llamafile@711344b
* fixed mul-mat error for old GPUs
* style fixes
* add mul mat src1 f16 test cases, fix more cases
ggml-ci
---------
Co-authored-by: bssrdf <bssrdf@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
The default values for tfs_z and typical_p were being set to zero, which
caused the token candidates array to get shrunk down to one element thus
preventing any sampling. Note this only applies to OpenAI API compatible
HTTP server requests.
The solution is to use the default values that OpenAI documents, as well
as ensuring we use the llama.cpp defaults for the rest. I've tested this
change still ensures deterministic output by default. If a "temperature"
greater than 0 is explicitly passed, then output is unique each time. If
"seed" is specified in addition to "temperature" then the output becomes
deterministic once more.
See mozilla-Ocho/llamafile#117
See mozilla-Ocho/llamafile@9e4bf29
* cuda : fix vmm pool with multi GPU
* hip
* use recommended granularity instead of minimum
* better error checking
* fix mixtral
* use cudaMemcpy3DPeerAsync
* use cuda_pool_alloc in ggml_cuda_op_mul_mat
* consolidate error checking in ggml_cuda_set_device
* remove unnecessary inlines
ggml-ci
* style fixes
* only use vmm for the main device
* fix scratch buffer size, re-enable vmm pool for all devices
* remove unnecessary check id != g_main_device
* cuda : improve cuda pool efficiency using virtual memory
* fix mixtral
* fix cmake build
* check for vmm support, disable for hip
ggml-ci
* fix hip build
* clarify granularity
* move all caps to g_device_caps
* refactor error checking
* add cuda_pool_alloc, refactor most pool allocations
ggml-ci
* fix hip build
* CUBLAS_TF32_TENSOR_OP_MATH is not a macro
* more hip crap
* llama : fix msvc warnings
* ggml : fix msvc warnings
* minor
* minor
* cuda : fallback to CPU on host buffer alloc fail
* Update ggml-cuda.cu
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* Update ggml-cuda.cu
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* ensure allocations are always aligned
* act_size -> actual_size
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* Check the full vocab for grammar only if necessary
* Fix missing logit restoration step (?)
Does this matter, actually?
* Fix whitespace / formatting
* Adjust comment
* Didn't mean to push test gbnf
* Split sampling into the helper function (?)
And also revert the changes made to the header
* common : fix final newline
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* initial commit, going through initializations
* main loop finished, starting to debug
* BUG: generates gibberish/repeating tokens after a while
* kv_cache management
* Added colors to distinguish drafted tokens (--color). Updated README
* lookup : fix token positions in the draft batch
* lookup : use n_draft from CLI params
* lookup : final touches
---------
Co-authored-by: Leon Ericsson <leon.ericsson@icloud.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* fix old jetson compile error
* Update Makefile
* update jetson detect and cuda version detect
* update cuda marco define
* update makefile and cuda,fix some issue
* Update README.md
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update Makefile
* Update README.md
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama : Add ability to cancel model load
Updated llama_progress_callback so that if it returns false, the model
loading is aborted.
* llama : Add test for model load cancellation
* Fix bool return in llama_model_load, remove std::ignore use
* Update llama.cpp
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Fail test if model file is missing
* Revert "Fail test if model file is missing"
This reverts commit 32ebd525bf.
* Add test-model-load-cancel to Makefile
* Revert "Revert "Fail test if model file is missing""
This reverts commit 2796953257.
* Simplify .gitignore for tests, clang-tidy fixes
* Label all ctest tests
* ci : ctest uses -L main
* Attempt at writing ctest_with_model
* ci : get ci/run.sh working with test-model-load-cancel
* ci : restrict .github/workflows/build.yml ctest to -L main
* update requirements.txt
* Disable test-model-load-cancel in make
* Remove venv before creation
* Restructure requirements.txt
Top-level now imports the specific additional requirements for each
python file. Using `pip install -r requirements.txt` will fail if
versions become mismatched in the per-file requirements.
* Make per-python-script requirements work alone
This doesn't break the main requirements.txt.
* Add comment
* Add convert-persimmon-to-gguf.py to new requirements.txt scheme
* Add check-requirements.sh script and GitHub workflow
* Remove shellcheck installation step from workflow
* Add nocleanup special arg
* Fix merge
see: https://github.com/ggerganov/llama.cpp/pull/4462#discussion_r1434593573
* reset to upstream/master
* Redo changes for cancelling model load
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* llama : initial ggml-backend integration
* add ggml-metal
* cuda backend can be used though ggml-backend with LLAMA_GGML_BACKEND_CUDA_TEST
access all tensor data with ggml_backend_tensor_get/set
* add ggml_backend_buffer_clear
zero-init KV cache buffer
* add ggml_backend_buffer_is_hos, used to avoid copies if possible when accesing tensor data
* disable gpu backends with ngl 0
* more accurate mlock
* unmap offloaded part of the model
* use posix_fadvise64(.., POSIX_FADV_SEQUENTIAL) to improve performance with mmap
* update quantize and lora
* update session copy/set to use ggml-backend
ggml-ci
* use posix_fadvise instead of posix_fadvise64
* ggml_backend_alloc_ctx_tensors_from_buft : remove old print
* llama_mmap::align_offset : use pointers instead of references for out parameters
* restore progress_callback behavior
* move final progress_callback call to load_all_data
* cuda : fix fprintf format string (minor)
* do not offload scales
* llama_mmap : avoid unmapping the same fragments again in the destructor
* remove unnecessary unmap
* metal : add default log function that prints to stderr, cleanup code
ggml-ci
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* allowed getting n_batch from llama_context in c api
* changed to use `uint32_t` instead of `int`
* changed to use `uint32_t` instead of `int` in `llama_n_ctx`
* Update llama.h
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* AMD ROCm: handle UMA memory VRAM expansions
This resolves#2797 by allowing ROCm AMD GPU users with a UMA to
dynamically expand the VRAM allocated to the GPU.
Without this, AMD ROCm users with shared CPU/GPU memory usually are
stuck with the BIOS-set (or fixed) framebuffer VRAM, making it
impossible to load more than 1-2 layers.
Note that the model is duplicated in RAM because it's loaded once for
the CPU and then copied into a second set of allocations that are
managed by the HIP UMA system. We can fix this later.
* clarify build process for ROCm on linux with cmake
* avoid using deprecated ROCm hipMallocHost
* keep simplifying the change required for UMA
* cmake: enable UMA-compatible allocation when LLAMA_HIP_UMA=ON
Otherwise, on Windows converting bling-phi-2-v0 (<https://huggingface.co/llmware/bling-phi-2-v0>) via convert-hf-to-gguf.py will fail with the following error:
```
Traceback (most recent call last):
File "C:\Users\User\git\gguf\convert-hf-to-gguf.py", line 1061, in <module>
model_instance.set_vocab()
File "C:\Users\User\git\gguf\convert-hf-to-gguf.py", line 52, in set_vocab
self._set_vocab_gpt2()
File "C:\Users\User\git\gguf\convert-hf-to-gguf.py", line 264, in _set_vocab_gpt2
special_vocab = gguf.SpecialVocab(dir_model, load_merges=True)
File "C:\Users\User\git\gguf\gguf\vocab.py", line 33, in __init__
self._load(Path(path))
File "C:\Users\User\git\gguf\gguf\vocab.py", line 81, in _load
self._try_load_merges_txt(path)
File "C:\Users\User\git\gguf\gguf\vocab.py", line 95, in _try_load_merges_txt
for line in fp:
File "C:\Users\User\miniconda3\envs\gguf\lib\encodings\cp1252.py", line 23, in decode
return codecs.charmap_decode(input,self.errors,decoding_table)[0]
UnicodeDecodeError: 'charmap' codec can't decode byte 0x81 in position 1415: character maps to <undefined>
```
regression of #4490
Adds defines for two new datatypes
cublasComputeType_t, cudaDataType_t.
Currently using deprecated hipblasDatatype_t since newer ones very recent.
* build : Check the ROCm installation location
* more generic approach
* fixup! It was returning the path instead of the command output
* fixup! Trailing whitespace
* Add API key authentication for enhanced server-client security
* server : to snake_case
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ggml : group mul_mat_id rows by matrix (cpu only)
* remove mmid parameters from mm forward
* store row groups in wdata and calculate only once in GGML_TASK_INIT
ggml-ci
* Fixes "Not enough space in the context's memory pool" encountered on certain models, which seems to be caused by some imprecision related to the automatic casting of floating point values
* do not cast to size_t, instead just use doubles
* ggml : add ggml_row_size(), deprecate ggml_type_sizef()
* ggml : fix row size compute to avoid overflows
* tests : fix sizey -> sizez
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* sync : ggml (SD ops, tests, kernels)
ggml-ci
* cuda : restore im2col
ggml-ci
* metal : fix accuracy of dequantization kernels
ggml-ci
* cuda : restore correct im2col
ggml-ci
* metal : try to fix moe test by reducing expert size
ggml-ci
* cuda : fix bin bcast when src1 and dst have different types
ggml-ci
---------
Co-authored-by: slaren <slarengh@gmail.com>
* convert : support Mixtral as LLAMA arch
* convert : fix n_ff typo
* llama : model loading
* ggml : sync latest ggml_mul_mat_id
* llama : update graph to support MoE
* llama : fix cur -> cur_expert
* llama : first working version
* llama : fix expert weighting in the FFN
* ggml : ggml_get_rows support 2D indexing [n_tokens, n_experts] (cpu only)
* ggml : add n_as argument to ggml_mul_mat_id
* ggml : fix ggml_get_rows to take into account ne02 / ne11
* metal : add more general support for ggml_get_rows + tests
* llama : add basic support for offloading moe with CUDA
* metal : add/mul/div use general kernel when src1 not cont
* metal : reduce the kernel launches for ggml_mul_mat_id
* ggml : get_rows : support non-contiguos tensors with gaps, generalize up to 3D
* ggml : update get_rows f16 and q
* cuda : support non-contiguous src1 in get_rows
* llama : offload missing ffn_moe_silu
* metal : fix ggml_get_rows to work with non-cont src1
* metal : add indirect mat-vec kernels for all quantization types
* llama : do not quantize expert gating tensors
* llama : add n_expert and n_expert_used to hparams + change quants
* test-backend-ops : add moe test
* cuda : fix get_rows when ncols is odd
* convert : determine n_ctx correctly
* metal : fix ggml_mul_mat_id for F32
* test-backend-ops : make experts more evenly probable (test_moe)
* test-backend-ops : cleanup, add moe test for batches
* test-backend-ops : add cpy from f32 -> all types test
* test-backend-ops : fix dequantize block offset
* llama : fix hard-coded number of experts
* test-backend-ops : simplify and disable slow tests to avoid CI timeout
* test-backend-ops : disable MOE test with thread sanitizer
* cuda : fix mul_mat_id with multi gpu
* convert : use 1e6 rope_freq_base for mixtral
* convert : fix style
* convert : support safetensors format
* gguf-py : bump version
* metal : add cpy f16 -> f32 kernel
* metal : fix binary ops for ne10 % 4 != 0
* test-backend-ops : add one more sum_rows test
* ggml : do not use BLAS with ggml_mul_mat_id
* convert-hf : support for mixtral-instruct (#4428)
* convert : typo fix, add additional hyperparameters, use LLaMA arch for Mixtral-instruct
* convert : use sentencepiece tokenizer for Mixtral-instruct
* convert : make flake8 happy
* metal : fix soft_max kernels
ref: 1914017863
* metal : limit kernels to not use more than the allowed threads
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Radek Pilar <github@mrkva.eu>
llama_context_params.logits_all is a parameter for controlling
llama_eval. This documents that logits_all should not be used with
llama_decode and llama_batch.
On commit b1108 (44c117f4) xaedes added
ggml_allocr * alloc = NULL;
... (many lines in between)
if (alloc) {
ggml_allocr_free(alloc);
}
Which is correct, but it's easy to lose context after many lines in between.
On commit b1287 (0e76a899) xaedes made a big change. From here on, alloc is freed eagerly.
alloc = ggml_allocr_new(...)
... (short lines of code)
ggml_allocr_free(alloc)
This happens a few times, but alloc is never set to NULL, and many lines below,
we still have
if (alloc) {
ggml_allocr_free(alloc);
}
which causes a double-free.
* reserve space for codepoints
* improvement for the appended 0
* used precomputed token text for grammar sample
* reserve canidates_decoded
* reserve canidates_grammar
* remove candidates_decoded
* Revert "remove candidates_decoded"
This reverts commit 3773328080.
* changed decode_utf8 to take src by ref
* feat: Allow overriding GGUF metadata when loading model
* Fix the one time GCC is stricter than clang about something
* Step1
* Refactor... basically everything!
* Nuke obsolete GetArrayLen struct
* simplify std::string specialization
* Various cleanups
Add informational output when overrides are applied
Warn user when an override with the wrong type is specified
* Fix broken logic for parsing bool KV overrides
Fix issue where overrides didn't apply when key missing in GGUF metadata
Resolve merge changes
* llama : rearrange model params
* Update new GET_KEY call
Add note that metadata KV overrides aren't reflected in initial metadata KV info dump
---------
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Samplers sequence order w parameter
* Cleaned commented code
* Fixed formatting
* Rewrote with unordered_map
* Revert and rewrite, too many problems and safeguards would be needed
* Fixed code style
* Code style fixes according to review
* More readable samplers input string, fixed help
* Style fix in sampler_queue
* Formatting fixes
* Fixing whitespaces
This commit updates the error message that is printed when the
KV cache is not big enough to hold all the prompt and generated
tokens. Specifically it removes the reference to n_parallel and
replaces it with n_len.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
This commit adds a requirements file for the convert-hf-to-gguf.py
script, and also add the torch and transformers packages to it.
The motivation for this is that currently running convert-hf-to-gguf.py
will produce the following error:
```console
$ python3 -m venv venv
$ source venv/bin/activate
(venv) $ pip install -r requirements.txt
Collecting numpy==1.24.4
Collecting sentencepiece==0.1.98
Collecting gguf>=0.1.0
Installing collected packages: sentencepiece, numpy, gguf
Successfully installed gguf-0.5.1 numpy-1.24.4 sentencepiece-0.1.98
(venv) $ python convert-hf-to-gguf.py --help
Traceback (most recent call last):
File "llama.cpp/convert-hf-to-gguf.py", line 16, in <module>
import torch
ModuleNotFoundError: No module named 'torch'
```
With this commit, and using requirements-hf-to-gguf.txt instead of
requirements.txt, the script can be run and shows the help output.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* metal : implement soft_max_ext
* cuda : implement soft_max_ext
* ggml : implement soft_max_ext (CPU)
* batched-bench : print threads
ggml-ci
* metal : simplify soft_max encoding
ggml-ci
* cuda : use 512 threads for soft_max instead of 32
* ggml : update soft max cpu
* cuda : do warp-based block reduce
* cuda : increase max block size to 1024
* cuda : fix warp reduction initialization of shared mem
* metal : warp-based reduction for soft max kernel
* metal : warp-based reduce for rms_norm
* metal : simplify soft max kernel
ggml-ci
* alloc : fix build with debug
* * add multiprompt support
* * cleanup
* * more cleanup
* * remove atomicity of id_gen, and change lock_guard to unique_lock on completion requests
* * remove all references to mutex_multitasks
* Update examples/server/server.cpp
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update examples/server/server.cpp
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update examples/server/server.cpp
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update examples/server/server.cpp
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* * change to set
---------
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* cmake : fix joining of REAL_GIT_DIR
* fix includes with help from include-what-you-use
* make : remove unneeded deps and add test-rope target
* fix C includes in C++ source files
* Revert "fix includes with help from include-what-you-use"
This reverts commit 635e9fadfd.
* ShareGPT4 compatibility (vision encoder only loading)
Load only a CLIP vision encoder (as supplied by ShareGPT finetunes)
Corrects the argument parsing for --img_mean and --img_std (which were previously not parsed but attempted to access)
Defines defaults for img_mean and img_std which are equal to the llava 1.5 CLIP encoder, so you do not have to provide them
* Update convert-image-encoder-to-gguf.py
* llama: fix alignment of general.name in print meta
This commit fixes the alignment of the general.name field in the
llm_load_print_meta function.
Currently the output looks like this:
```console
llm_load_print_meta: model ftype = mostly Q4_0
llm_load_print_meta: model params = 13.02 B
llm_load_print_meta: model size = 6.86 GiB (4.53 BPW)
llm_load_print_meta: general.name = LLaMA v2
```
And with this commit it looks like this:
```console
llm_load_print_meta: model ftype = mostly Q4_0
llm_load_print_meta: model params = 13.02 B
llm_load_print_meta: model size = 6.86 GiB (4.53 BPW)
llm_load_print_meta: general.name = LLaMA v2
```
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* llama: fix alignment of special tokens
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
Typical sampling was broken because after copying new_candidates into canditates, the "sorted" bool is left at "true", but the new data is no longer sorted according to probability. Patch to set "sorted" to false.
Test: Generating with temp=0.0001 (approx. argmax) should generate the same sequence at typical>=1.0 and typical=0.9999 (approx. disabled, but enters the typical sampling codepath).
* fix oai proxy
fix generation not stoped while bot stop talking in chat mode
fix possible `slot_id` not exist
response for cors (and pre flight)
* oai proxy: workaround for some client (such as Chatbox)
* use stop as separator to replace hardcoded `\n`
* ggml : use blas even if src0 is not F32
* llama : use n_threads_batch only when n_tokens >= 32
ggml-ci
* llama : revert n_threads_batch logic
ggml-ci
* copy to llama.cpp as subdir
* attempt enabling metal, fails
* ggml metal compiles!
* Update README.md
* initial conversion to new format, utf8 errors?
* bug fixes, but now has an invalid memory access :(
* added O3, now has insufficient memory access
* begin sync with master
* update to match latest code, new errors
* fixed it!
* fix for loop conditionals, increase result size
* fix current workflow errors
* attempt a llama.swiftui workflow
* Update .github/workflows/build.yml
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Add openai-compatible POST /v1/chat/completions API endpoint to server example
* fix code style
* Update server README.md
* Improve server README.md
* Fix server.cpp code style according to review
* server : some style changes
* server : indentation
* server : enable special tokens during tokenization by default
* server : minor code style
* server : change random string generator
* straightforward /v1/models endpoint
---------
Co-authored-by: kir-gadjello <111190790+kir-gadjello@users.noreply.github.com>
Co-authored-by: Tobi Lütke <tobi@Tobis-MacBook-Pro.local>
* llama : keep track of used KV cells + better KV cache management
* llama : zero KV cache used upon clear
ggml-ci
* llama : allow exporting a view of the KV cache (#4180)
* Allow exporting a view of the KV cache
* Allow dumping the sequences per cell in common
* Track max contiguous cells value and position as well
* Fix max contiguous empty cells index calculation
Make dump functions deal with lengths or sequences counts > 10 better
* Fix off by one error in dump_kv_cache_view
* Add doc comments for KV cache view functions
Eliminate cell sequence struct; use llama_seq_id directly
Minor cleanups
* common : add -dkvc arg for enabling kv cache dumps
---------
Co-authored-by: Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com>
Disabled rules:
* E203 Whitespace before ':' - disabled because we often use 'C' Style where values are aligned
* E211 Whitespace before '(' (E211) - disabled because we often use 'C' Style where values are aligned
* E221 Multiple spaces before operator - disabled because we often use 'C' Style where values are aligned
* E225 Missing whitespace around operator - disabled because it's broken so often it seems like a standard
* E231 Missing whitespace after ',', ';', or ':' - disabled because we often use 'C' Style where values are aligned
* E241 Multiple spaces after ',' - disabled because we often use 'C' Style where values are aligned
* E251 Unexpected spaces around keyword / parameter equals - disabled because it's broken so often it seems like a standard
* E261 At least two spaces before inline comment - disabled because it's broken so often it seems like a standard
* E266 Too many leading '#' for block comment - sometimes used as "section" separator
* E501 Line too long - disabled because it's broken so often it seems like a standard
* E701 Multiple statements on one line (colon) - broken only in convert.py when defining abstract methods (we can use# noqa instead)
* E704 Multiple statements on one line - broken only in convert.py when defining abstract methods (we can use# noqa instead)
* Support special tokens and not adding BOS to prompt in speculative
* Adapt to new should_add_bos function
* Ensure tgt and dft have same add_bos setting
* ggml-cuda.cu: Clean up warnings when compiling with clang
* ggml-cuda.cu: Move static items into anonymous namespace
* ggml-cuda.cu: Fix use of namespace start macro
* Revert "ggml-cuda.cu: Fix use of namespace start macro"
This reverts commit 26c1149026.
* Revert "ggml-cuda.cu: Move static items into anonymous namespace"
This reverts commit e29757e0f7.
* build: support ppc64le build for make and CMake
* build: keep __POWER9_VECTOR__ ifdef and extend with __powerpc64__
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
- introduces help entry for the argument
- cuts '--gpu-layers' form in order to simplify usage and documentation.
Signed-off-by: Jiri Podivin <jpodivin@gmail.com>
Co-authored-by: Jiri Podivin <jpodivin@redhat.com>
* Remove logically superfluous assertions and order by dimension
* Use cblas_sgemm() to implement ggml_compute_forward_out_prod()
* Remove ggml_compute_forward_out_prod_use_blas(), fix compiling errors on cmake/zig, remove trailing whitespace
* Add openBLAS support for sgemm() in compute_forward_out_prod()
* finetune : zero the loraB initial vectors
Without this, the first iteration is starting out far from the base model, instead of exactly on it.
Zeroing loraB is what the paper recommends. loralib also zeroes at least one of the init vector pairs
(though it departs from the paper in using a different distribution for the other vector, in some cases).
* tabs to spaces
* Use ggml_set_zero instead of adding a new function
* gguf-py: gguf-dump: Respect --no-tensor flag in JSON mode.
* Respect add_bos_token GGUF metadata value
* gguf-py: Try to fix SpecialVocab giving up too easily for the Nth time
* add safetensors to convert.py help message
* Check for single-file safetensors model
* Update convert.py "model" option help message
* revert convert.py help message change
* gguf-py: Refactor and add file reading support
* Replay changes from #3871
Credit to @cebtenzzre for that pull
* Various type annotation fixes.
* sort imports with isort (again)
* Fix missing return statement in add_tensor
* style cleanup with flake8
* fix NamedTuple and Enum usage
* Fix an issue with state init in GGUFReader
Move examples to an examples/ directory
Clean up examples
Add an example of modifying keys in a GGUF file
Update documentation with info on examples
Try to support people importing gguf/gguf.py directly
* Damagage is not a word.
* Clean up gguf-py/examples/modify_gguf.py whitespace
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update gguf-py/examples/modify_gguf.py formatting
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update gguf-py/gguf/gguf_reader.py type hint
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Make examples executable, formatting changes
* Add more information to GGUFReader and examples comments
* Include a gguf Python package version bump
* Add convert-gguf-endian.py script
* cleanup
* gguf-py : bump minor version
* Reorganize scripts
* Make GGUFReader endian detection less arbitrary
* Add JSON dumping support to gguf-dump.py
Which I kind of regret now
* A few for gguf-dump.py cleanups
* Murder accidental tuple in gguf-py/scripts/gguf-dump.py
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* cleanup
* constants : remove unneeded type annotations
* fix python 3.8 compat
* Set up gguf- scripts in pyproject.toml
* And include scripts/__init__.py, derp
* convert.py: We can't currently support Q8_0 on big endian.
* gguf-py: SpecialVocab: Always try available sources for special token ids
gguf-py: SpecialVocab: Try to load merges from merges.txt if not in tokenizer.json
gguf-py: SpecialVocab: Add 'add_bos_token' type bools to GGUF metadata
u
* cleanup
* Promote add_X_token to GGUF metadata for BOS and EOS
---------
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update server.cpp with min_p after it was introduced in https://github.com/ggerganov/llama.cpp/pull/3841
* Use spaces instead of tabs
* Update index.html.hpp after running deps.sh
* Fix test - fix line ending
* fix backward process of rope
rope backward process was broken after YaRN RoPE (#2268) implementation, due to missing changes in backward functions.
the code for the backward process is nearly identically to the forward process:
the only difference is the sign of the sin-values.
to avoid future regressions remove the near-duplicate backward functions and reuse the forward code:
for this a new function argument `bool forward` was added to `ggml_compute_forward_rope_f32` and `ggml_compute_forward_rope_f16`.
the sin-values will be negated when forward is false.
* fix finetune rope call to use correct default attn_factor of 1.0f
* remove unused `ggml_rope_xpos_back`
it is better to have only one `ggml_rope_back` function that accepts all rope parameters, so that `ggml_compute_backward` can propagate all parameters without having to switch between different rope_back variants.
* fix comments explaining the sinus sign in ggml_forward_rope
* add missing function arguments in declaration
* fix function argument type in declaration
llava-cli was loading models with default params and ignoring settings
from the cli. This switches to a generic function to load the params
from the cli options.
* wip llava python bindings compatibility
* add external llava API
* add base64 in-prompt image support
* wip refactor image loading
* refactor image load out of llava init
* cleanup
* further cleanup; move llava-cli into its own file and rename
* move base64.hpp into common/
* collapse clip and llava libraries
* move llava into its own subdir
* wip
* fix bug where base64 string was not removed from the prompt
* get libllava to output in the right place
* expose llava methods in libllama.dylib
* cleanup memory usage around clip_image_*
* cleanup and refactor *again*
* update headerdoc
* build with cmake, not tested (WIP)
* Editorconfig
* Editorconfig
* Build with make
* Build with make
* Fix cyclical depts on Windows
* attempt to fix build on Windows
* attempt to fix build on Windows
* Upd TODOs
* attempt to fix build on Windows+CUDA
* Revert changes in cmake
* Fix according to review comments
* Support building as a shared library
* address review comments
---------
Co-authored-by: M. Yusuf Sarıgöz <yusufsarigoz@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
* Add detection code for avx
* Only check hardware when option is ON
* Modify per code review sugguestions
* Build locally will detect CPU
* Fixes CMake style to use lowercase like everywhere else
* cleanup
* fix merge
* linux/gcc version for testing
* msvc combines avx2 and fma into /arch:AVX2 so check for both
* cleanup
* msvc only version
* style
* Update FindSIMD.cmake
---------
Co-authored-by: Howard Su <howard0su@gmail.com>
Co-authored-by: Jeremy Dunn <jeremydunn123@gmail.com>
* Revert "cuda : add ROCM aliases for CUDA pool stuff (#3918)"
This reverts commit 629f917cd6.
* Revert "cuda : use CUDA memory pool with async memory allocation/deallocation when available (#3903)"
This reverts commit d6069051de.
ggml-ci
* Using cuda memory pools for async alloc/dealloc.
* If cuda device doesnt support memory pool than use old implementation.
* Removed redundant cublasSetStream
---------
Co-authored-by: Oleksii Maryshchenko <omaryshchenko@dtis.com>
* cmake : fix build when .git does not exist
* cmake : simplify BUILD_INFO target
* cmake : add missing dependencies on BUILD_INFO
* build : link against build info instead of compiling against it
* zig : make build info a .cpp source instead of a header
Co-authored-by: Matheus C. França <matheus-catarino@hotmail.com>
* cmake : revert change to CMP0115
---------
Co-authored-by: Matheus C. França <matheus-catarino@hotmail.com>
* Add '-ngl' support to finetune.cpp
* Add fprintf in ggml_cuda_op_add
When I tried CUDA offloading during finetuning following the readme, I got an assert here.
This probably isn't an important case because inference later gives a warning saying you should use f16 or f32 instead when using lora
* Add 'finetune.sh', which currently fails when using GPU
"error: operator (): Finetuning on tensors with type 'f16' is not yet supported"
* tweak finetune.sh
* Suppress some warnings in ggml.c
* Add f16 implementation to ggml_compute_forward_add_f16_f32
* Add an f16 case to ggml_add_cast_impl and llama_build_lora_finetune_graphs
* finetune.sh: Edit comments
* Add "add_f16_f32_f32_cuda"
* Tweak an error message
* finetune.sh: Add an optional LLAMA_MODEL_DIR variable
* finetune.sh: Add an optional LLAMA_TRAINING_DIR variable
* train : minor
* tabs to spaces
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
* Introduce the new Min-P sampler by @kalomaze
The Min-P sampling method was designed as an alternative to Top-P, and aims to ensure a balance of quality and variety. The parameter *p* represents the minimum probability for a token to be considered, relative to the probability of the most likely token.
* Min-P enabled and set to 0.05 default
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
* Extend llama_kv_cache_seq_rm to allow matichng any sequence
* Replace llama_kv_cache_tokens_rm with llama_kv_cache_clear
Use llama_kv_cache_clear for cache clearing
Change calls to llama_kv_cache_tokens_rm that want to delete by position to use llama_kv_cache_seq_rm functionality
* Try cwd for ggml-metal if bundle lookup fails
When building with `-DBUILD_SHARED_LIBS=ON -DLLAMA_METAL=ON -DLLAMA_BUILD_SERVER=ON`,
`server` would fail to load `ggml-metal.metal` because `[bundle pathForResource:...]`
returns `nil`. In that case, fall back to `ggml-metal.metal` in the cwd instead of
passing `null` as a path.
Follows up on #1782
* Update ggml-metal.m
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Allow quantizing k-quants to fall back when tensor size incompatible
* quantizing: Add warning when tensors were incompatible with k-quants
Clean up k-quants state passing a bit
* cuda : prints wip
* cuda : new cublas gemm branch for multi-batch quantized src0
* cuda : add F32 sgemm branch
* cuda : fine-tune >= VOLTA params + use MMQ only for small batches
* cuda : remove duplicated cuBLAS GEMM code
* cuda : add CUDA_USE_TENSOR_CORES and GGML_CUDA_FORCE_MMQ macros
* build : add compile option to force use of MMQ kernels
* cmake : add helper for faster CUDA builds
* batched : add NGL arg
* ggml : skip nops in compute_forward
* cuda : minor indentation
* cuda : batched cuBLAS GEMMs for src0 F16 and src1 F32 (attention ops)
* Apply suggestions from code review
These changes plus:
```c++
#define cublasGemmBatchedEx hipblasGemmBatchedEx
```
are needed to compile with ROCM. I haven't done performance testing, but it seems to work.
I couldn't figure out how to propose a change for lines outside what the pull changed, also this is the first time trying to create a multi-part review so please forgive me if I mess something up.
* cuda : add ROCm / hipBLAS cublasGemmBatchedEx define
* cuda : add cublasGemmStridedBatchedEx for non-broadcasted cases
* cuda : reduce mallocs in cublasGemmBatchedEx branch
* cuda : add TODO for calling cublas from kernel + using mem pool
---------
Co-authored-by: Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com>
We still have the heads up in `README.md` regarding `bpe` tokenizers and this patch is needed for
- a couple of tokenizer tests
- some more `special` and `non-special` added tokens handling (as far as I understand it)
* Update special token handling
* Add mpt
* added `llama_model_token_*` variants to all the `llama_token_*` functions.
* added `LLAMA_API`
* formatting
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* removed old `llama_token` functions
* changed 3 more functions to take in model
- `llama_token_get_text`
- `llama_token_get_score`
- `llama_token_get_type`
* added back docs
* fixed main.cpp
* changed token functions to use new model variants
* changed token functions to use new model variants
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* implementing parallel decoding in server example
* crash fixed
* save dev progress
* refactored sampling function
* completion endpoint working
* multiple client support
* grammar + no stream completion
* cached prompt support
* chat.mjs support cached prompt + some fixes
* server ui now support multiple clients
* unused change reverted
* fixed timings per slot
* add context swap
* add changes to README.md
* llava multimodal integration
* fixed tokens probs
* add multimodal input - alfa
* refactor code + remove unused comments + improved README.md
* fix compilation errors with llvm
* notify the user from server ui that multimodality is unavialable
* some ci fixes
* fix ci make build undefined ref errors
* fix long prompt than ctx proposed in #3639
* fixed premature end due stop word
* context shift fixed
* fix llava implementation
* sync README.md changes
* readme change
* update api like OpenAI
* multimodal support enabled by default
* fix make bui;d errors
* fix multiple clients
* fix zig build
* new sampling API
* latest changes of sampling API
* server : coding-style normalization
* server : coding-style normalization (part 2)
* server : remove beam-search functionality
* server : bug fix in ingest_images
n_tokens is incremented internally by llama_batch_add
* server : use refs + use llama_batch_clear()
* server : snake case
* server : minor sync
* added thread safe pipeline
* server : bach has to be allocated for n_parallel sequences
* server : no need for atomic int - already using mutex
* server : logs + minor code style
* server : fix multibyte handle in partial response (#3706)
* fix image load + view image in chat
* make : silence stb warnings
* clip : link to ggml, not to llama
* server : fix switch fallthrough
* server : fix crash in Debug on macOS (I have no idea why this fixes it!?)
* server : refactor ctx_sampling init + n_ctx + names
* server : bug fix for prompt caching
* Do not save/load image_data to localStorage
* editorconfig : new line in index.html
* server : completion requests remember slot_id
* Update readme to document multimodal in server
* server : minor style
* Update readme to document multimodal in server
* server : hide ctx_sampling->prev behind API (#3696)
* server : apply fix from #3722
* server : fix slot reuse
* server : add comment about changing slot_state to bool
---------
Co-authored-by: FSSRepo <go778sgt@gmail.com>
Co-authored-by: Damian Stewart <d@damianstewart.com>
Co-authored-by: Steward Garcia <57494570+FSSRepo@users.noreply.github.com>
Co-authored-by: Jhen-Jie Hong <iainst0409@gmail.com>
Co-authored-by: M. Yusuf Sarıgöz <yusufsarigoz@gmail.com>
* Add validation for special token ids to llama.cpp
Small optimization for llama_byte_to_token SPM mode
* Fix BPE newline check, only I could break something so simple
* Killll meeeeee
* Account for GGUF_KEY_KEY only setting when the key exists
* Minor code cleanups.
* Fix convert.py error msg when added tokens are out of range
* Make gguf SpecialVocab vocab size-aware
Update conversion scripts accordingly
* Avoid a string copy
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* infill tokens correction
* serverinfill tokens correction
* removing any leading whitespace from infill suffix and removing leeading space token from suffix when params.escape
* removing any leading whitespace from infill suffix and removing leeading space token from suffix when params.escape
* only rm when params.escape, rm space if possible which is added back or rm added space token
* only rm when params.escape, rm space if possible which is added back or rm added space token
* Revert "only rm when params.escape, rm space if possible which is added back or rm added space token"
This reverts commit 63ba0b621f.
* fix interactive prompt escaping and fix server infill leading space handling
* rm unnecessary bool check
* process escapes for neg prompt and interactive consec prompts
* removed unneccessary static string escape
* check whether platform is 390x if yes->do not import immintrin.h
* support s390x big endian
* support --bigendian option for s390x
1. verified with baichuan7b-chat with float 16 on s390x
2. verified with baichuan7b-chat
3. verified with chinese-alpaca-2-13b-f16
* update format based on editor-config checker result
* Update convert-baichuan-hf-to-gguf.py
* 1. check in ggml.c if endianess is not match
2. update GGUF version
3. change get_pack_prefix to property
4. update information log
* always use "GGUF" as beginng of GGUF file
* Compare "GGUF" with file header char by char
1. Set GGUF_MAGIC to "GGUF" string instead of int value
2. Compare "GGUF" char by char to ensure its byte order
3. Move bytes swap code from convert.py to gguf.py write_tensor_data
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* metal : implement dequantize_q5_0
* metal : block_q_n_dot_y for block_q5_0 (broken)
* metal : revert unnecessary change
* metal : implement dequantize_q5_1
* metal : block_q_n_dot_y for q5_1 (broken)
* metal : fix block_q_n_dot_y
* minor : spaces / formatting
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Added documentation of JSON return value of /completion endpoint
* Update examples/server/README.md
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Rewrite special token handling from #1931
* shorten param name, add st verification by type
* use offsets instead of copy by substr
* formatting, remove copying iterator on delete
* llama : normalize code-style
* swift fix
* print pfx/sfx if verb, main: split pfx input sfx
* dont add space when using special tokens
* minor : comment + spacing
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit removes `n_threads` from the `llama_decode_internal`
functions doc comment as it does not exist anymore.
It looks like this parameter was removed in
Commit 16bc66d947 ("llama.cpp : split
llama_context_params into model and context params").
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* WIP: start implementing LLaVA
* rm scratch buf for now, will revert after cleanup
* LLaVA image encoder is working. will combine with llama
* Add llava inference code, but it's buggy. debugging
* LLaVA is working e2e, needs to optimize memory allocation + cleanup
* Use ggml_allocr + rm unnecessary code
* fix: crlf -> lf
* fix: new line at EoF
* fix: trailing whitespace
* Add readme
* Update readme
* Some cleanup
* Are you happy editorconfig?
* rm unused batch image preprocessing
* rm unused import
* fix: rm designated initializers
* introduce pad-to-square mode for non-square images
* are you happy editorconfig?
* gitignore /llava
* Handle cases where image file does not exist
* add llava target to Makefile
* add support for 13b model variant
* Maybe seed is unlucky?
* Check if apples are compared to apples
* are you happy editorconfig?
* Use temperature = 0.1 by default
* command line: use gpt_params_parse()
* minor
* handle default n_predict
* fix typo
* llava : code formatting, rename files, fix compile warnings
* do not use Wno-cast-qual for MSVC
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Fix mirostat state when using multiple sequences
* Fix mirostat by completely refactoring sampling!
* Try to fix zig build.
* Export function to fetch/create default sampler states
Code formatting cleanups and add some comments
Silence a warning about id not being used when logging is disabled
* Apply some renaming suggestions.
Fix comments that were out of sync with the pull.
* Use more consistant naming convention for sampling contexts
* CUDA: added support for ggml_clamp (see also: https://github.com/ggerganov/ggml/issues/545)
* mpt : added an implementation based (mostly) on falcon integration, modified with deltas from ggml/examples/mpt
* mpt : protect against "clip_qkv": null in mpt-7b
* mpt : quick fix to avoid "Strange model" warning when quantizing MPT models
* mpt : addendum to changeset:84e30e8 - leave parameter clamp_kqv out from metadata rather than use 0.0 to indicate "no clamping" (more compliant with the current GGUF spec?)
* mpt : standardized all tensor names to follow GGUF spec
* mpt : addendum to changeset:1be89c40 - use "req" parameter of GGUF_GET_KEY macro instead of duplicate code
* mpt : fixed comment s/gptneox/mpt/
* mpt : remove tabs, trailing whitespace
* mpt : removed ne01 + n_past == ne00 assertion from alibi (cuda/f32) and rope_shift from build_mpt
* mpt : updated convert-mpt-hf-to-gguf.py to reflect changes made to convert-gptneox-hf-to-gguf.py in pr:3252
* comment out n_past instead of marking it unused
* mpt : removed hardcoded +178 from convert script in favor of utilizing hparams["vocab_size"]
* mpt : remove unused tokenizer_json in convert script
* ggml : remove obsolete n_past assert in ggml_alibi
* llama : print clam_kqv and max_alibi_bias hparams
---------
Co-authored-by: Cebtenzzre <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* infill tokens correction
* serverinfill tokens correction
* removing any leading whitespace from infill suffix and removing leeading space token from suffix when params.escape
* removing any leading whitespace from infill suffix and removing leeading space token from suffix when params.escape
* only rm when params.escape, rm space if possible which is added back or rm added space token
* only rm when params.escape, rm space if possible which is added back or rm added space token
* Revert "only rm when params.escape, rm space if possible which is added back or rm added space token"
This reverts commit 63ba0b621f.
* fix interactive prompt escaping and fix server infill leading space handling
* rm unnecessary bool check
Please include information about your system, the steps to reproduce the bug, and the version of llama.cpp that you are using. If possible, please provide a minimal code example that reproduces the bug.
If the bug concerns the server, please try to reproduce it first using the [server test scenario framework](https://github.com/ggerganov/llama.cpp/tree/master/examples/server/tests).
about: Used to report issues and request enhancements for llama.cpp
title: "[User] Insert summary of your issue or enhancement.."
labels: ''
assignees: ''
---
# Prerequisites
Please answer the following questions for yourself before submitting an issue.
- [ ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
- [ ] I carefully followed the [README.md](https://github.com/ggerganov/llama.cpp/blob/master/README.md).
- [ ] I [searched using keywords relevant to my issue](https://docs.github.com/en/issues/tracking-your-work-with-issues/filtering-and-searching-issues-and-pull-requests) to make sure that I am creating a new issue that is not already open (or closed).
- [ ] I reviewed the [Discussions](https://github.com/ggerganov/llama.cpp/discussions), and have a new bug or useful enhancement to share.
# Expected Behavior
Please provide a detailed written description of what you were trying to do, and what you expected `llama.cpp` to do.
# Current Behavior
Please provide a detailed written description of what `llama.cpp` did, instead.
# Environment and Context
Please provide detailed information about your computer setup. This is important in case the issue is not reproducible except for under certain specific conditions.
* Physical (or virtual) hardware you are using, e.g. for Linux:
`$ lscpu`
* Operating System, e.g. for Linux:
`$ uname -a`
* SDK version, e.g. for Linux:
```
$ python3 --version
$ make --version
$ g++ --version
```
# Failure Information (for bugs)
Please help provide information about the failure if this is a bug. If it is not a bug, please remove the rest of this template.
# Steps to Reproduce
Please provide detailed steps for reproducing the issue. We are not sitting in front of your screen, so the more detail the better.
1. step 1
2. step 2
3. step 3
4. etc.
# Failure Logs
Please include any relevant log snippets or files. If it works under one configuration but not under another, please provide logs for both configurations and their corresponding outputs so it is easy to see where behavior changes.
Also, please try to **avoid using screenshots** if at all possible. Instead, copy/paste the console output and use [Github's markdown](https://docs.github.com/en/get-started/writing-on-github/getting-started-with-writing-and-formatting-on-github/basic-writing-and-formatting-syntax) to cleanly format your logs for easy readability.
Example environment info:
```
llama.cpp$ git log | head -1
commit 2af23d30434a677c6416812eea52ccc0af65119c
llama.cpp$ lscpu | egrep "AMD|Flags"
Vendor ID: AuthenticAMD
Model name: AMD Ryzen Threadripper 1950X 16-Core Processor
Please close your issue when it has been answered.
@duncan-donut: I'm trying to figure out what kind of "support" you need for this script and why, exactly? Is there a question about how the code works that hasn't already been addressed in one or more comments below this ticket, or are we talking something else entirely like some sorta bugfixing job because your server setup is different from mine??
I can understand if your site needs to be running smoothly and you need help with a fix of sorts but there should really be nothing wrong here that the code itself could not handle. And given that I'm getting reports about how it works perfectly well on some other servers, what exactly are we talking? A detailed report will do wonders in helping us get this resolved for ya quickly so please take your time and describe the issue(s) you see as clearly & concisely as possible!!
@duncan-donut: I'm not sure if you have access to cPanel but you could try these instructions. It is worth a shot! Let me know how it goes (or what error message, exactly!) when/if ya give that code a go? [end of text]
main: mem per token = 71159620 bytes
main: load time = 19309.95 ms
main: sample time = 168.62 ms
main: predict time = 223895.61 ms / 888.47 ms per token
main: total time = 246406.42 ms
Performance counter stats for './main -m ./models/65B/ggml-model-q4_0.bin -t 16 -n 1024 -p Please close your issue when it has been answered.':
Please answer the following questions for yourself before submitting an issue.
- [ ] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
- [ ] I carefully followed the [README.md](https://github.com/ggerganov/llama.cpp/blob/master/README.md).
- [ ] I [searched using keywords relevant to my issue](https://docs.github.com/en/issues/tracking-your-work-with-issues/filtering-and-searching-issues-and-pull-requests) to make sure that I am creating a new issue that is not already open (or closed).
- [ ] I reviewed the [Discussions](https://github.com/ggerganov/llama.cpp/discussions), and have a new bug or useful enhancement to share.
# Feature Description
Please provide a detailed written description of what you were trying to do, and what you expected `llama.cpp` to do as an enhancement.
# Motivation
Please provide a detailed written description of reasons why this feature is necessary and how it is useful to `llama.cpp` users.
# Possible Implementation
If you have an idea as to how it can be implemented, please write a detailed description. Feel free to give links to external sources or share visuals that might be helpful to understand the details better.
SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators—such as CPUs, GPUs, and FPGAs. It is a single-source embedded domain-specific language based on pure C++17.
oneAPI is a specification that is open and standards-based, supporting multiple architecture types including but not limited to GPU, CPU, and FPGA. The spec has both direct programming and API-based programming paradigms.
Intel uses the SYCL as direct programming language to support CPU, GPUs and FPGAs.
To avoid to re-invent the wheel, this code refer other code paths in llama.cpp (like OpenBLAS, cuBLAS, CLBlast). We use a open-source tool [SYCLomatic](https://github.com/oneapi-src/SYCLomatic) (Commercial release [Intel® DPC++ Compatibility Tool](https://www.intel.com/content/www/us/en/developer/tools/oneapi/dpc-compatibility-tool.html)) migrate to SYCL.
The llama.cpp for SYCL is used to support Intel GPUs.
For Intel CPU, recommend to use llama.cpp for X86 (Intel MKL building).
## News
- 2024.3
- Support multiple cards: **--split-mode**: [none|layer]; not support [row], it's on developing.
- Support to assign main GPU by **--main-gpu**, replace $GGML_SYCL_DEVICE.
- Support detecting all GPUs with level-zero and same top **Max compute units**.
|Intel iGPU| Support| iGPU in i5-1250P, i7-1260P, i7-1165G7|
Note: If the EUs (Execution Unit) in iGPU is less than 80, the inference speed will be too slow to use.
### Memory
The memory is a limitation to run LLM on GPUs.
When run llama.cpp, there is print log to show the applied memory on GPU. You could know how much memory to be used in your case. Like `llm_load_tensors: buffer size = 3577.56 MiB`.
For iGPU, please make sure the shared memory from host memory is enough. For llama-2-7b.Q4_0, recommend the host memory is 8GB+.
For dGPU, please make sure the device memory is enough. For llama-2-7b.Q4_0, recommend the device memory is 4GB+.
## Nvidia GPU
### Verified
|Intel GPU| Status | Verified Model|
|-|-|-|
|Ampere Series| Support| A100|
### oneMKL
The current oneMKL release does not contain the oneMKL cuBlas backend.
As a result for Nvidia GPU's oneMKL must be built from source.
a. Please follow the procedure in [Get the Intel® oneAPI Base Toolkit ](https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit.html).
Recommend to install to default folder: **/opt/intel/oneapi**.
Following guide use the default folder as example. If you use other folder, please modify the following guide info with your folder.
b. Check
```sh
source /opt/intel/oneapi/setvars.sh
sycl-ls
```
There should be one or more level-zero devices. Please confirm that at least one GPU is present, like **[ext_oneapi_level_zero:gpu:0]**.
max compute_units 512, max work group size 1024, max sub group size 32, global mem size 16225243136
```
|Attribute|Note|
|-|-|
|compute capability 1.3|Level-zero running time, recommended |
|compute capability 3.0|OpenCL running time, slower than level-zero in most cases|
4. Set device ID and execute llama.cpp
Set device ID = 0 by **GGML_SYCL_DEVICE=0**
```sh
GGML_SYCL_DEVICE=0 ./build/bin/main -m models/llama-2-7b.Q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 400 -e -ngl 33
```
or run by script:
```sh
./examples/sycl/run_llama2.sh
```
Note:
- By default, mmap is used to read model file. In some cases, it leads to the hang issue. Recommend to use parameter **--no-mmap** to disable mmap() to skip this issue.
5. Check the device ID in output
Like:
```
Using device **0** (Intel(R) Arc(TM) A770 Graphics) as main device
```
## Windows
### Setup Environment
1. Install Intel GPU driver.
Please install Intel GPU driver by official guide: [Install GPU Drivers](https://www.intel.com/content/www/us/en/products/docs/discrete-gpus/arc/software/drivers.html).
Note: **The driver is mandatory for compute function**.
2. Install Visual Studio.
Please install [Visual Studio](https://visualstudio.microsoft.com/) which impact oneAPI environment enabling in Windows.
3. Install Intel® oneAPI Base toolkit.
a. Please follow the procedure in [Get the Intel® oneAPI Base Toolkit ](https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit.html).
Recommend to install to default folder: **C:\Program Files (x86)\Intel\oneAPI**.
Following guide uses the default folder as example. If you use other folder, please modify the following guide info with your folder.
b. Enable oneAPI running environment:
- In Search, input 'oneAPI'.
Search & open "Intel oneAPI command prompt for Intel 64 for Visual Studio 2022"
max compute_units 512, max work group size 1024, max sub group size 32, global mem size 16225243136
```
|Attribute|Note|
|-|-|
|compute capability 1.3|Level-zero running time, recommended |
|compute capability 3.0|OpenCL running time, slower than level-zero in most cases|
4. Set device ID and execute llama.cpp
Set device ID = 0 by **set GGML_SYCL_DEVICE=0**
```
set GGML_SYCL_DEVICE=0
build\bin\main.exe -m models\llama-2-7b.Q4_0.gguf -p "Building a website can be done in 10 simple steps:\nStep 1:" -n 400 -e -ngl 33 -s 0
```
or run by script:
```
.\examples\sycl\win-run-llama2.bat
```
Note:
- By default, mmap is used to read model file. In some cases, it leads to the hang issue. Recommend to use parameter **--no-mmap** to disable mmap() to skip this issue.
5. Check the device ID in output
Like:
```
Using device **0** (Intel(R) Arc(TM) A770 Graphics) as main device
```
## Environment Variable
#### Build
|Name|Value|Function|
|-|-|-|
|LLAMA_SYCL|ON (mandatory)|Enable build with SYCL code path. <br>For FP32/FP16, LLAMA_SYCL=ON is mandatory.|
|LLAMA_SYCL_F16|ON (optional)|Enable FP16 build with SYCL code path. Faster for long-prompt inference. <br>For FP32, not set it.|
|CMAKE_C_COMPILER|icx|Use icx compiler for SYCL code path|
|CMAKE_CXX_COMPILER|icpx (Linux), icx (Windows)|use icpx/icx for SYCL code path|
#### Running
|Name|Value|Function|
|-|-|-|
|GGML_SYCL_DEVICE|0 (default) or 1|Set the device id used. Check the device ids by default running output|
|GGML_SYCL_DEBUG|0 (default) or 1|Enable log function by macro: GGML_SYCL_DEBUG|
|ZES_ENABLE_SYSMAN| 0 (default) or 1|Support to get free memory of GPU by sycl::aspect::ext_intel_free_memory.<br>Recommended to use when --split-mode = layer|
## Known Issue
- Hang during startup
llama.cpp use mmap as default way to read model file and copy to GPU. In some system, memcpy will be abnormal and block.
Solution: add **--no-mmap** or **--mmap 0**.
- Split-mode: [row] is not supported
It's on developing.
## Q&A
- Error: `error while loading shared libraries: libsycl.so.7: cannot open shared object file: No such file or directory`.
Miss to enable oneAPI running environment.
Install oneAPI base toolkit and enable it by: `source /opt/intel/oneapi/setvars.sh`.
- In Windows, no result, not error.
Miss to enable oneAPI running environment.
- Meet compile error.
Remove folder **build** and try again.
- I can **not** see **[ext_oneapi_level_zero:gpu:0]** afer install GPU driver in Linux.
Please run **sudo sycl-ls**.
If you see it in result, please add video/render group to your ID:
```
sudo usermod -aG render username
sudo usermod -aG video username
```
Then **relogin**.
If you do not see it, please check the installation GPU steps again.
Inference of [LLaMA](https://arxiv.org/abs/2302.13971) model in pure C/C++
Inference of Meta's [LLaMA](https://arxiv.org/abs/2302.13971) model (and others) in pure C/C++
### Recent API changes
- [2024 Mar 8] `llama_kv_cache_seq_rm()` returns a `bool` instead of `void`, and new `llama_n_max_seq()` returns the upper limit of acceptable `seq_id` in batches (relevant when dealing with multiple sequences) https://github.com/ggerganov/llama.cpp/pull/5328
- [2024 Mar 4] Embeddings API updated https://github.com/ggerganov/llama.cpp/pull/5796
- [2024 Mar 3] `struct llama_context_params` https://github.com/ggerganov/llama.cpp/pull/5849
### Hot topics
-‼️ Breaking change: `rope_freq_base` and `rope_freq_scale` must be set to zero to use the model default values: [#3401](https://github.com/ggerganov/llama.cpp/pull/3401)
-Parallel decoding + continuous batching support added: [#3228](https://github.com/ggerganov/llama.cpp/pull/3228) \
<li><a href="#obtaining-the-facebook-llama-original-model-and-stanford-alpaca-model-data">Obtaining the Facebook LLaMA original model and Stanford Alpaca model data</a></li>
<li><a href="#verifying-the-model-files">Verifying the model files</a></li>
- [X] [Baichuan-7B](https://huggingface.co/baichuan-inc/baichuan-7B) and its derivations (such as [baichuan-7b-sft](https://huggingface.co/hiyouga/baichuan-7b-sft))
[llama.cpp web server](./examples/server) is a lightweight [OpenAI API](https://github.com/openai/openai-openapi) compatible HTTP server that can be used to serve local models and easily connect them to existing clients.
Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently three different implementations of it:
Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Support with CPU-only BLAS implementations doesn't affect the normal generation performance. We may see generation performance improvements with GPU-involved BLAS implementations, e.g. cuBLAS, hipBLAS and CLBlast. There are currently several different BLAS implementations available for build and use:
- #### Accelerate Framework:
@@ -365,20 +416,37 @@ Building the program with BLAS support may lead to some performance improvements
Check [BLIS.md](docs/BLIS.md) for more information.
- #### Intel MKL
- #### SYCL
SYCL is a higher-level programming model to improve programming productivity on various hardware accelerators.
By default, `LLAMA_BLAS_VENDOR` is set to `Generic`, so if you already sourced intel environment script and assign `-DLLAMA_BLAS=ON` in cmake, the mkl version of Blas will automatically been selected. You may also specify it by:
llama.cpp based on SYCL is used to **support Intel GPU** (Data Center Max series, Flex series, Arc series, Built-in GPU and iGPU).
For detailed info, please refer to [llama.cpp for SYCL](README-sycl.md).
- #### Intel oneMKL
Building through oneAPI compilers will make avx_vnni instruction set available for intel processors that do not support avx512 and avx512_vnni. Please note that this build config **does not support Intel GPU**. For Intel GPU support, please refer to [llama.cpp for SYCL](./README-sycl.md).
- Using manual oneAPI installation:
By default, `LLAMA_BLAS_VENDOR` is set to `Generic`, so if you already sourced intel environment script and assign `-DLLAMA_BLAS=ON` in cmake, the mkl version of Blas will automatically been selected. Otherwise please install oneAPI and follow the below steps:
```bash
mkdir build
cd build
source /opt/intel/oneapi/setvars.sh # You can skip this step if in oneapi-basekit docker image, only required for manual installation
If you do not want to source the environment vars and install oneAPI manually, you can also build the code using intel docker container: [oneAPI-basekit](https://hub.docker.com/r/intel/oneapi-basekit). Then, you can use the commands given above.
Check [Optimizing and Running LLaMA2 on Intel® CPU](https://www.intel.com/content/www/us/en/content-details/791610/optimizing-and-running-llama2-on-intel-cpu.html) for more information.
- #### cuBLAS
This provides BLAS acceleration using the CUDA cores of your Nvidia GPU. Make sure to have the CUDA toolkit installed. You can download it from your Linux distro's package manager (e.g. `apt install nvidia-cuda-toolkit`) or from here: [CUDA Toolkit](https://developer.nvidia.com/cuda-downloads).
For Jetson user, if you have Jetson Orin, you can try this: [Offical Support](https://www.jetson-ai-lab.com/tutorial_text-generation.html). If you are using an old model(nano/TX2), need some additional operations before compiling.
- Using `make`:
```bash
make LLAMA_CUBLAS=1
@@ -411,22 +479,39 @@ Building the program with BLAS support may lead to some performance improvements
This provides BLAS acceleration on HIP-supported AMD GPUs.
Make sure to have ROCm installed.
You can download it from your Linux distro's package manager or from here: [ROCm Quick Start (Linux)](https://rocm.docs.amd.com/en/latest/deploy/linux/quick_start.html).
Windows support is coming soon...
- Using `make`:
```bash
make LLAMA_HIPBLAS=1
```
- Using `CMake`:
- Using `CMake` for Linux (assuming a gfx1030-compatible AMD GPU):
On Linux it is also possible to use unified memory architecture (UMA) to share main memory between the CPU and integrated GPU by setting `-DLLAMA_HIP_UMA=ON"`.
However, this hurts performance for non-integrated GPUs (but enables working with integrated GPUs).
- Using `make` (example for target gfx1030, build with 16 CPU threads):
```bash
make -j16 LLAMA_HIPBLAS=1 LLAMA_HIP_UMA=1 AMDGPU_TARGETS=gxf1030
```
- Using `CMake` for Windows (using x64 Native Tools Command Prompt for VS, and assuming a gfx1100-compatible AMD GPU):
Make sure that `AMDGPU_TARGETS` is set to the GPU arch you want to compile for. The above example uses `gfx1100` that corresponds to Radeon RX 7900XTX/XT/GRE. You can find a list of targets [here](https://llvm.org/docs/AMDGPUUsage.html#processors)
Find your gpu version string by matching the most significant version information from `rocminfo | grep gfx | head -1 | awk '{print $2}'` with the list of processors, e.g. `gfx1035` maps to `gfx1030`.
The environment variable [`HIP_VISIBLE_DEVICES`](https://rocm.docs.amd.com/en/latest/understand/gpu_isolation.html#hip-visible-devices) can be used to specify which GPU(s) will be used.
If your GPU is not officialy supported you can use the environment variable [`HSA_OVERRIDE_GFX_VERSION`] set to a similar GPU, for example 10.3.0 on RDNA2 or 11.0.0 on RDNA3.
If your GPU is not officially supported you can use the environment variable [`HSA_OVERRIDE_GFX_VERSION`] set to a similar GPU, for example 10.3.0 on RDNA2 (e.g. gfx1030, gfx1031, or gfx1035) or 11.0.0 on RDNA3.
The following compilation options are also available to tweak performance (yes, they refer to CUDA, not HIP, because it uses the same code as the cuBLAS version above):
| Option | Legal values | Default | Description |
@@ -539,43 +624,108 @@ Building the program with BLAS support may lead to some performance improvements
You can get a list of platforms and devices from the `clinfo -l` command, etc.
### Prepare Data & Run
- #### Vulkan
**With docker**:
You don't need to install Vulkan SDK. It will be installed inside the container.
# To verify the installation, use the command below:
vulkaninfo
```
Alternatively your package manager might be able to provide the appropiate libraries. For example for Ubuntu 22.04 you can install `libvulkan-dev` instead.
Then, build llama.cpp using the cmake command below:
```bash
mkdir -p build
cd build
cmake .. -DLLAMA_VULKAN=1
cmake --build . --config Release
# Test the output binary (with "-ngl 33" to offload all layers to GPU)
./bin/main -m "PATH_TO_MODEL" -p "Hi you how are you" -n 50 -e -ngl 33 -t 4
# You should see in the output, ggml_vulkan detected your GPU. For example:
To obtain the official LLaMA 2 weights please see the <a href="#obtaining-and-using-the-facebook-llama-2-model">Obtaining and using the Facebook LLaMA 2 model</a> section. There is also a large selection of pre-quantized `gguf` models available on Hugging Face.
```bash
# obtain the original LLaMA model weights and place them in ./models
# obtain the official LLaMA model weights and place them in ./models
When running the larger models, make sure you have enough disk space to store all the intermediate files.
### Running on Windows with prebuilt binaries
You will find prebuilt Windows binaries on the release page.
Simply download and extract the latest zip package of choice: (e.g. `llama-b1380-bin-win-avx2-x64.zip`)
From the unzipped folder, open a terminal/cmd window here and place a pre-converted `.gguf` model file. Test out the main example like so:
```
.\main -m llama-2-7b.Q4_0.gguf -n 128
```
### Memory/Disk Requirements
As the models are currently fully loaded into memory, you will need adequate disk space to save them and sufficient RAM to load them. At the moment, memory and disk requirements are the same.
| Model | Original size | Quantized size (4-bit) |
2. Run `./perplexity -m models/7B/ggml-model-q4_0.gguf -f wiki.test.raw`
3. Output:
```
@@ -629,7 +791,7 @@ And after 4.45 hours, you will have the final perplexity.
### Interactive mode
If you want a more ChatGPT-like experience, you can run in interactive mode by passing `-i` as a parameter.
In this mode, you can always interrupt generation by pressing Ctrl+C and entering one or more lines of text, which will be converted into tokens and appended to the current context. You can also specify a *reverse prompt* with the parameter `-r "reverse prompt string"`. This will result in user input being prompted whenever the exact tokens of the reverse prompt string are encountered in the generation. A typical use is to use a prompt that makes LLaMa emulate a chat between multiple users, say Alice and Bob, and pass `-r "Alice:"`.
In this mode, you can always interrupt generation by pressing Ctrl+C and entering one or more lines of text, which will be converted into tokens and appended to the current context. You can also specify a *reverse prompt* with the parameter `-r "reverse prompt string"`. This will result in user input being prompted whenever the exact tokens of the reverse prompt string are encountered in the generation. A typical use is to use a prompt that makes LLaMA emulate a chat between multiple users, say Alice and Bob, and pass `-r "Alice:"`.
Here is an example of a few-shot interaction, invoked with the command
@@ -679,9 +841,9 @@ The `grammars/` folder contains a handful of sample grammars. To write your own,
For authoring more complex JSON grammars, you can also check out https://grammar.intrinsiclabs.ai/, a browser app that lets you write TypeScript interfaces which it compiles to GBNF grammars that you can save for local use. Note that the app is built and maintained by members of the community, please file any issues or FRs on [its repo](http://github.com/intrinsiclabsai/gbnfgen) and not this one.
### Instruction mode with Alpaca
### Instruct mode
1. First, download the `ggml` Alpaca model into the `./models` folder
1. First, download and place the `ggml` model into the `./models` folder
2. Run the `main` tool like this:
```
@@ -693,7 +855,7 @@ Sample run:
```
== Running in interactive mode. ==
- Press Ctrl+C to interject at any time.
- Press Return to return control to LLaMa.
- Press Return to return control to LLaMA.
- If you want to submit another line, end your input in '\'.
Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Using [OpenLLaMA](https://github.com/openlm-research/open_llama)
OpenLLaMA is an openly licensed reproduction of Meta's original LLaMA model. It uses the same architecture and is a drop-in replacement for the original LLaMA weights.
- Download the [3B](https://huggingface.co/openlm-research/open_llama_3b), [7B](https://huggingface.co/openlm-research/open_llama_7b), or [13B](https://huggingface.co/openlm-research/open_llama_13b) model from Hugging Face.
- Convert the model to ggml FP16 format using `python convert.py <path to OpenLLaMA directory>`
### Using [GPT4All](https://github.com/nomic-ai/gpt4all)
*Note: these instructions are likely obsoleted by the GGUF update*
- Obtain the `tokenizer.model` file from LLaMA model and put it to `models`
- Obtain the `added_tokens.json` file from Alpaca model and put it to `models`
- Obtain the `gpt4all-lora-quantized.bin` file from GPT4All model and put it to `models/gpt4all-7B`
- It is distributed in the old `ggml` format which is now obsoleted
- You have to convert it to the new format using `convert.py`:
- You can now use the newly generated `models/gpt4all-7B/ggml-model-q4_0.bin` model in exactly the same way as all other models
- The newer GPT4All-J model is not yet supported!
### Using Pygmalion 7B & Metharme 7B
- Obtain the [LLaMA weights](#obtaining-the-facebook-llama-original-model-and-stanford-alpaca-model-data)
- Obtain the [Pygmalion 7B](https://huggingface.co/PygmalionAI/pygmalion-7b/) or [Metharme 7B](https://huggingface.co/PygmalionAI/metharme-7b) XOR encoded weights
- Convert the LLaMA model with [the latest HF convert script](https://github.com/huggingface/transformers/blob/main/src/transformers/models/llama/convert_llama_weights_to_hf.py)
- Merge the XOR files with the converted LLaMA weights by running the [xor_codec](https://huggingface.co/PygmalionAI/pygmalion-7b/blob/main/xor_codec.py) script
- Convert to `ggml` format using the `convert.py` script in this repo:
```bash
python3 convert.py pygmalion-7b/ --outtype q4_1
```
> The Pygmalion 7B & Metharme 7B weights are saved in [bfloat16](https://en.wikipedia.org/wiki/Bfloat16_floating-point_format) precision. If you wish to convert to `ggml` without quantizating, please specify the `--outtype` as `f32` instead of `f16`.
### Obtaining the Facebook LLaMA original model and Stanford Alpaca model data
- **Under no circumstances should IPFS, magnet links, or any other links to model downloads be shared anywhere in this repository, including in issues, discussions, or pull requests. They will be immediately deleted.**
- The LLaMA models are officially distributed by Facebook and will **never** be provided through this repository.
- Refer to [Facebook's LLaMA repository](https://github.com/facebookresearch/llama/pull/73/files) if you need to request access to the model data.
### Obtaining and using the Facebook LLaMA 2 model
- Refer to [Facebook's LLaMA download page](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) if you want to access the model data.
Please verify the [sha256 checksums](SHA256SUMS) of all downloaded model files to confirm that you have the correct model data files before creating an issue relating to your model files.
- The following python script will verify if you have all possible latest files in your self-installed `./models` subdirectory:
```bash
# run the verification script
./scripts/verify-checksum-models.py
```
- On linux or macOS it is also possible to run the following commands to verify if you have all possible latest files in your self-installed `./models` subdirectory:
- On Linux: `sha256sum --ignore-missing -c SHA256SUMS`
- on macOS: `shasum -a 256 --ignore-missing -c SHA256SUMS`
### Seminal papers and background on the models
If your issue is with model generation quality, then please at least scan the following links and papers to understand the limitations of LLaMA models. This is especially important when choosing an appropriate model size and appreciating both the significant and subtle differences between LLaMA models and ChatGPT:
@@ -860,19 +964,22 @@ Place your desired model into the `~/llama.cpp/models/` directory and execute th
* Create a folder to store big models & intermediate files (ex. /llama/models)
#### Images
We have two Docker images available for this project:
We have three Docker images available for this project:
1. `ghcr.io/ggerganov/llama.cpp:full`: This image includes both the main executable file and the tools to convert LLaMA models into ggml and convert into 4-bit quantization. (platforms: `linux/amd64`, `linux/arm64`)
2. `ghcr.io/ggerganov/llama.cpp:light`: This image only includes the main executable file. (platforms: `linux/amd64`, `linux/arm64`)
3. `ghcr.io/ggerganov/llama.cpp:server`: This image only includes the server executable file. (platforms: `linux/amd64`, `linux/arm64`)
Additionally, there the following images, similar to the above:
- `ghcr.io/ggerganov/llama.cpp:full-cuda`: Same as `full` but compiled with CUDA support. (platforms: `linux/amd64`)
- `ghcr.io/ggerganov/llama.cpp:light-cuda`: Same as `light` but compiled with CUDA support. (platforms: `linux/amd64`)
- `ghcr.io/ggerganov/llama.cpp:server-cuda`: Same as `server` but compiled with CUDA support. (platforms: `linux/amd64`)
- `ghcr.io/ggerganov/llama.cpp:full-rocm`: Same as `full` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`)
- `ghcr.io/ggerganov/llama.cpp:light-rocm`: Same as `light` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`)
- `ghcr.io/ggerganov/llama.cpp:server-rocm`: Same as `server` but compiled with ROCm support. (platforms: `linux/amd64`, `linux/arm64`)
The GPU enabled images are not currently tested by CI beyond being built. They are not built with any variation from the ones in the Dockerfiles defined in [.devops/](.devops/) and the Gitlab Action defined in [.github/workflows/docker.yml](.github/workflows/docker.yml). If you need different settings (for example, a different CUDA or ROCm library, you'll need to build the images locally for now).
The GPU enabled images are not currently tested by CI beyond being built. They are not built with any variation from the ones in the Dockerfiles defined in [.devops/](.devops/) and the GitHub Action defined in [.github/workflows/docker.yml](.github/workflows/docker.yml). If you need different settings (for example, a different CUDA or ROCm library, you'll need to build the images locally for now).
#### Usage
@@ -896,6 +1003,12 @@ or with a light image:
docker run -v /path/to/models:/models ghcr.io/ggerganov/llama.cpp:light -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512
Assuming one has the [nvidia-container-toolkit](https://github.com/NVIDIA/nvidia-container-toolkit) properly installed on Linux, or is using a GPU enabled cloud, `cuBLAS` should be accessible inside the container.
@@ -905,6 +1018,7 @@ Assuming one has the [nvidia-container-toolkit](https://github.com/NVIDIA/nvidia
You may want to pass in some different `ARGS`, depending on the CUDA environment supported by your container host, as well as the GPU architecture.
@@ -918,6 +1032,7 @@ The resulting images, are essentially the same as the non-CUDA images:
1. `local/llama.cpp:full-cuda`: This image includes both the main executable file and the tools to convert LLaMA models into ggml and convert into 4-bit quantization.
2. `local/llama.cpp:light-cuda`: This image only includes the main executable file.
3. `local/llama.cpp:server-cuda`: This image only includes the server executable file.
#### Usage
@@ -926,6 +1041,7 @@ After building locally, Usage is similar to the non-CUDA examples, but you'll ne
```bash
docker run --gpus all -v /path/to/models:/models local/llama.cpp:full-cuda --run -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 --n-gpu-layers 1
docker run --gpus all -v /path/to/models:/models local/llama.cpp:light-cuda -m /models/7B/ggml-model-q4_0.gguf -p "Building a website can be done in 10 simple steps:" -n 512 --n-gpu-layers 1
docker run --gpus all -v /path/to/models:/models local/llama.cpp:server-cuda -m /models/7B/ggml-model-q4_0.gguf --port 8000 --host 0.0.0.0 -n 512 --n-gpu-layers 1
```
### Contributing
@@ -945,12 +1061,13 @@ docker run --gpus all -v /path/to/models:/models local/llama.cpp:light-cuda -m /
- There are no strict rules for the code style, but try to follow the patterns in the code (indentation, spaces, etc.). Vertical alignment makes things more readable and easier to batch edit
- Clean-up any trailing whitespaces, use 4 spaces for indentation, brackets on the same line, `void * ptr`, `int & a`
- See [good first issues](https://github.com/ggerganov/llama.cpp/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) for tasks suitable for first contributions
- Tensors store data in row-major order. We refer to dimension 0 as columns, 1 as rows, 2 as matrices
- Matrix multiplication is unconventional: [`z = ggml_mul_mat(ctx, x, y)`](https://github.com/ggerganov/llama.cpp/blob/880e352277fc017df4d5794f0c21c44e1eae2b84/ggml.h#L1058-L1064) means `zT = x @ yT`
message(WARNING"Git index not found in git repository.")
set(GIT_INDEX"")
endif()
else()
message(WARNING"Git repository not found; to enable automatic generation of build info, make sure Git is installed and the project is a Git repository.")
set(GIT_INDEX"")
endif()
# Add a custom command to rebuild build-info.cpp when .git/index changes
fprintf(stderr," --sample-start STR Sets the starting point for samples after the specified pattern. If empty use every token position as sample start. (default '%s')\n",params->sample_start.c_str());
fprintf(stderr," --include-sample-start Include the sample start in the samples. (default off)\n");
fprintf(stderr," --overlapping-samples Samples my overlap, will include sample-start of second and following samples. When off, samples will end at begin of next sample. (default off)\n");
fprintf(stderr," --overlapping-samples Samples may overlap, will include sample-start of second and following samples. When off, samples will end at begin of next sample. (default off)\n");
fprintf(stderr," --fill-with-next-samples Samples shorter than context length will be followed by the next (shuffled) samples. (default off)\n");
fprintf(stderr," --separate-with-eos When fill-with-next-samples, insert end-of-sequence token between samples.%s\n",params->separate_with_eos?" (default)":"");
fprintf(stderr," --separate-with-bos When fill-with-next-samples, insert begin-of-sequence token between samples.%s\n",params->separate_with_bos?" (default)":"");
If you see these lines, then the GPU is being used.
## Verifying that the CPU is not oversaturated
llama accepts a `-t N` (or `--threads N`) parameter. It's extremely important that this parameter is not too large. If your token generation is extremely slow, try setting this number to 1. If this significantly improves your token generation speed, then your CPU is being oversaturated and you need to explicitly set this parameter to the number of the physicial CPU cores on your machine (even if you utilize a GPU). If in doubt, start with 1 and double the amount until you hit a performance bottleneck, then scale the number down.
llama accepts a `-t N` (or `--threads N`) parameter. It's extremely important that this parameter is not too large. If your token generation is extremely slow, try setting this number to 1. If this significantly improves your token generation speed, then your CPU is being oversaturated and you need to explicitly set this parameter to the number of the physical CPU cores on your machine (even if you utilize a GPU). If in doubt, start with 1 and double the amount until you hit a performance bottleneck, then scale the number down.
# Example of runtime flags effect on inference speed benchmark
## [LLaVA](https://github.com/haotian-liu/LLaVA/) example (llava.py)
1. Obtian LLaVA model (following https://github.com/haotian-liu/LLaVA/ , use https://huggingface.co/liuhaotian/LLaVA-13b-delta-v1-1/).
2. Convert it to ggml format.
3.`llava_projection.pth` is [pytorch_model-00003-of-00003.bin](https://huggingface.co/liuhaotian/LLaVA-13b-delta-v1-1/blob/main/pytorch_model-00003-of-00003.bin).
torch.save({k: dic[k] for k in used_key}, pth_path)
```
4. Check the path of LLaVA model and `llava_projection.pth` in `llava.py`.
## [PandaGPT](https://github.com/yxuansu/PandaGPT) example (panda_gpt.py)
1. Obtian PandaGPT lora model from https://github.com/yxuansu/PandaGPT. Rename the file to `adapter_model.bin`. Use [convert-lora-to-ggml.py](../../convert-lora-to-ggml.py) to convert it to ggml format.
"what is the text in the picture? 'llama' or 'lambda'?")
a.chat("what is the color of it?")
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
