flake8 fix

convert_hf_to_gguf.py

@@ -3733,13 +3733,6 @@ class Ernie4_5Model(TextModel):
     def set_vocab(self):
         self._set_vocab_sentencepiece()
 
-        tokenizer_config_file = self.dir_model / 'tokenizer_config.json'
-        if tokenizer_config_file.is_file():
-            with open(tokenizer_config_file, "r", encoding="utf-8") as f:
-                tokenizer_config_json = json.load(f)
-                if "add_prefix_space" in tokenizer_config_json:
-                    self.gguf_writer.add_add_space_prefix(tokenizer_config_json["add_prefix_space"])
-
     def set_gguf_parameters(self):
         super().set_gguf_parameters()
 
@@ -3749,10 +3742,6 @@ class Ernie4_5Model(TextModel):
         if (head_dim := self.hparams.get("head_dim")) is None:
             head_dim = self.hparams["hidden_size"] // num_heads
 
-        if "mlp_AR" in name or "vision_model" in name:
-            # skip vision model and projector tensors
-            return
-
         if "ernie." in name:
             name = name.replace("ernie.", "model.")
         # split the qkv weights
@@ -3862,48 +3851,6 @@ class Ernie4_5MoeModel(Ernie4_5Model):
                 raise ValueError(f"Unprocessed experts: {experts}")
 
 
-@ModelBase.register("PaddleOCRVLForConditionalGeneration")
-class PaddleOCRModel(Ernie4_5Model):
-    model_arch = gguf.MODEL_ARCH.PADDLEOCR
-
-
-@ModelBase.register("PaddleOCRVisionModel")
-class PaddleOCRVisionModel(MmprojModel):
-    # PaddleOCR-VL uses a modified version of Siglip
-    min_pixels: int = 0
-    max_pixels: int = 0
-
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        assert self.hparams_vision is not None
-        self.min_pixels = self.preprocessor_config["min_pixels"]
-        self.max_pixels = self.preprocessor_config["max_pixels"]
-        self.hparams_vision["image_size"] = int(math.sqrt(self.max_pixels))
-
-    def set_gguf_parameters(self):
-        super().set_gguf_parameters()
-        assert self.hparams_vision is not None
-        hparams = self.hparams_vision
-        self.gguf_writer.add_clip_projector_type(gguf.VisionProjectorType.PADDLEOCR)
-        self.gguf_writer.add_vision_max_pixels(self.max_pixels)
-        self.gguf_writer.add_vision_min_pixels(self.min_pixels)
-        self.gguf_writer.add_vision_use_gelu(True)
-        self.gguf_writer.add_vision_attention_layernorm_eps(hparams.get("rms_norm_eps", 1e-6))
-
-    def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
-        name = name.replace("visual.", "model.")
-
-        if "vision_model" in name or "mlp_AR" in name:
-            if "packing_position_embedding" in name:
-                return # unused
-            elif "vision_model.head" in name:
-                # we don't yet support image embeddings for this model
-                return
-            else:
-                yield from super().modify_tensors(data_torch, name, bid)
-        return # skip other tensors
-
-
 @ModelBase.register(
     "Qwen2VLModel",
     "Qwen2VLForConditionalGeneration",

docs/ops.md

@@ -31,7 +31,7 @@ Legend:
 |                          CONV_3D | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
 |                CONV_TRANSPOSE_1D | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                CONV_TRANSPOSE_2D | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |
-|                              COS | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
+|                              COS | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ❌ | ❌ | ❌ |
 |                      COUNT_EQUAL | ❌ | ✅ | ✅ | ✅ | ✅ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                              CPY | ❌ | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | 🟡 | ❌ | ❌ |
 |               CROSS_ENTROPY_LOSS | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ | ❌ |
@@ -96,13 +96,13 @@ Legend:
 |                          SIGMOID | ❌ | ✅ | ✅ | 🟡 | 🟡 | 🟡 | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                             SILU | ❌ | ✅ | ✅ | 🟡 | 🟡 | 🟡 | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                        SILU_BACK | ❌ | ❌ | ✅ | ✅ | ❌ | ❌ | ❌ | ✅ | ❌ | ❌ | ❌ |
-|                              SIN | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
+|                              SIN | ❌ | ✅ | ✅ | ✅ | 🟡 | ❌ | ✅ | 🟡 | ❌ | ❌ | ❌ |
 |                         SOFTPLUS | ❌ | ❌ | ✅ | 🟡 | 🟡 | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |
 |                         SOFT_MAX | ❌ | 🟡 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ |
 |                    SOFT_MAX_BACK | ❌ | ❌ | 🟡 | 🟡 | ❌ | ❌ | 🟡 | ✅ | ❌ | ❌ | ❌ |
 |                        SOLVE_TRI | ❌ | ❌ | ✅ | 🟡 | ❌ | ❌ | ❌ | 🟡 | ❌ | ❌ | ❌ |
-|                              SQR | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ✅ | ❌ | ❌ |
-|                             SQRT | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ✅ | ❌ | ❌ |
+|                              SQR | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ❌ | ❌ | ❌ |
+|                             SQRT | ❌ | ✅ | ✅ | ✅ | 🟡 | ✅ | ✅ | 🟡 | ❌ | ❌ | ❌ |
 |                         SSM_CONV | ❌ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |
 |                         SSM_SCAN | ❌ | ❌ | ✅ | ✅ | ✅ | ❌ | ❌ | 🟡 | ❌ | ❌ | ❌ |
 |                             STEP | ❌ | ✅ | ✅ | 🟡 | 🟡 | ❌ | ✅ | 🟡 | ✅ | ❌ | ❌ |

docs/ops/WebGPU.csv

@@ -8760,14 +8760,22 @@
 "WebGPU: WebGPU","ADD_ID","type_a=f32,type_b=f32,n_embd=129,n_experts=8,n_experts_used=4,n_token=1","support","0","no","WebGPU"
 "WebGPU: WebGPU","ADD_ID","type_a=f32,type_b=f32,n_embd=129,n_experts=8,n_experts_used=4,n_token=32","support","0","no","WebGPU"
 "WebGPU: WebGPU","ADD_ID","type_a=f32,type_b=f32,n_embd=129,n_experts=8,n_experts_used=4,n_token=129","support","0","no","WebGPU"
+"WebGPU: WebGPU","SQR","type=f16,ne=[10,5,4,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SQRT","type=f16,ne=[10,3,3,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","LOG","type=f16,ne=[10,5,4,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SIN","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","COS","type=f16,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CLAMP","type=f16,ne=[10,5,4,3],min=-0.500000,max=0.500000","support","1","yes","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f16,ne_a=[10,5,4,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","CEIL","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SQR","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SQRT","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","LOG","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SIN","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","COS","type=f16,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CLAMP","type=f16,ne=[7,1,5,3],min=-0.500000,max=0.500000","support","1","yes","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f16,ne_a=[7,1,5,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
@@ -8778,14 +8786,22 @@
 "WebGPU: WebGPU","ROUND","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SQR","type=f32,ne=[10,5,4,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SQRT","type=f32,ne=[10,3,3,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","LOG","type=f32,ne=[10,5,4,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SIN","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
+"WebGPU: WebGPU","COS","type=f32,ne=[10,2,2,2]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CLAMP","type=f32,ne=[10,5,4,3],min=-0.500000,max=0.500000","support","1","yes","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f32,ne_a=[10,5,4,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","CEIL","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","ROUND","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
 "WebGPU: WebGPU","TRUNC","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SQR","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","SQRT","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","LOG","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
+"WebGPU: WebGPU","SIN","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
+"WebGPU: WebGPU","COS","type=f32,ne=[7,1,5,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","CLAMP","type=f32,ne=[7,1,5,3],min=-0.500000,max=0.500000","support","1","yes","WebGPU"
 "WebGPU: WebGPU","LEAKY_RELU","type=f32,ne_a=[7,1,5,3],negative_slope=0.100000","support","0","no","WebGPU"
 "WebGPU: WebGPU","FLOOR","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
@@ -18885,27 +18901,3 @@
 "WebGPU: WebGPU","CROSS_ENTROPY_LOSS_BACK","type=f32,ne=[30000,1,1,1]","support","0","no","WebGPU"
 "WebGPU: WebGPU","OPT_STEP_ADAMW","type=f32,ne=[10,5,4,3]","support","0","no","WebGPU"
 "WebGPU: WebGPU","OPT_STEP_SGD","type=f32,ne=[10,5,4,3]","support","0","no","WebGPU"
-"WebGPU: WebGPU","SQR","type=f16,ne=[10,5,4,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f16,ne=[10,3,3,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f16,ne=[10,2,2,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQR","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQR","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f16,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f16,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQR","type=f32,ne=[10,5,4,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f32,ne=[10,3,3,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f32,ne=[10,2,2,2]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQR","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQR","type=f32,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SQRT","type=f32,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","SIN","type=f32,ne=[1024,1024,1,1]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f32,ne=[7,1,5,3]","support","1","yes","WebGPU"
-"WebGPU: WebGPU","COS","type=f32,ne=[1024,1024,1,1]","support","1","yes","WebGPU"

ggml/include/ggml.h

@@ -730,10 +730,6 @@ extern "C" {
     GGML_API size_t  ggml_type_size(enum ggml_type type);             // size in bytes for all elements in a block
     GGML_API size_t  ggml_row_size (enum ggml_type type, int64_t ne); // size in bytes for all elements in a row
 
-    GGML_DEPRECATED(
-    GGML_API double ggml_type_sizef(enum ggml_type type), // ggml_type_size()/ggml_blck_size() as float
-    "use ggml_row_size() instead");
-
     GGML_API const char * ggml_type_name(enum ggml_type type);
     GGML_API const char * ggml_op_name  (enum ggml_op   op);
     GGML_API const char * ggml_op_symbol(enum ggml_op   op);

ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp

@@ -57,8 +57,6 @@ layout (push_constant) uniform parameter
     uint nbi1;
     uint ne11;
 #else
-    uint base_work_group_z;
-    uint num_batches;
     uint k_split;
     uint ne02;
     uint ne12;
@@ -110,7 +108,7 @@ void main() {
     const uint ic = gl_WorkGroupID.y;
 
 #ifdef MUL_MAT_ID
-    const uint expert_idx = gl_WorkGroupID.z;
+    const uint expert_idx = gl_GlobalInvocationID.z;
     if (ic * BN >= data_expert_count[expert_idx]) {
         return;
     }
@@ -120,7 +118,7 @@ void main() {
 #endif
 
 #ifndef MUL_MAT_ID
-    const uint batch_idx = gl_WorkGroupID.z + p.base_work_group_z;
+    const uint batch_idx = gl_GlobalInvocationID.z;
 
     const uint i13 = batch_idx / p.ne12;
     const uint i12 = batch_idx % p.ne12;
@@ -278,7 +276,7 @@ void main() {
     const uint dc = ic * BN + warp_c * WN;
 
 #ifndef MUL_MAT_ID
-    const uint offsets = batch_idx * p.batch_stride_d + ik * p.batch_stride_d * p.num_batches;
+    const uint offsets = batch_idx * p.batch_stride_d + ik * p.batch_stride_d * gl_NumWorkGroups.z;
 #endif
 
     [[unroll]] for (uint wsic = 0; wsic < WNITER; wsic++) {

ggml/src/ggml-webgpu/ggml-webgpu.cpp

@@ -2008,14 +2008,6 @@ static std::optional<webgpu_command> ggml_webgpu_encode_node(webgpu_context ctx,
             return ggml_webgpu_unary_op(ctx, src0, node);
         case GGML_OP_LOG:
             return ggml_webgpu_unary_op(ctx, src0, node);
-        case GGML_OP_SQR:
-            return ggml_webgpu_unary_op(ctx, src0, node);
-        case GGML_OP_SQRT:
-            return ggml_webgpu_unary_op(ctx, src0, node);
-        case GGML_OP_SIN:
-            return ggml_webgpu_unary_op(ctx, src0, node);
-        case GGML_OP_COS:
-            return ggml_webgpu_unary_op(ctx, src0, node);
         case GGML_OP_PAD:
             return ggml_webgpu_pad(ctx, src0, node);
         case GGML_OP_ARGMAX:
@@ -2975,18 +2967,6 @@ static bool ggml_backend_webgpu_device_supports_op(ggml_backend_dev_t dev, const
         case GGML_OP_LOG:
             supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
             break;
-        case GGML_OP_SQR:
-            supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
-            break;
-        case GGML_OP_SQRT:
-            supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
-            break;
-        case GGML_OP_SIN:
-            supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
-            break;
-        case GGML_OP_COS:
-            supports_op = (op->type == GGML_TYPE_F32 || op->type == GGML_TYPE_F16) && (src0->type == op->type);
-            break;
         case GGML_OP_PAD:
             supports_op = op->type == GGML_TYPE_F32 && src0->type == GGML_TYPE_F32;
             break;

ggml/src/ggml-webgpu/wgsl-shaders/unary.wgsl

@@ -170,20 +170,6 @@ fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
 #ifdef TRUNC
     let res = trunc(src[params.offset_src + src_idx]);
 #endif
-#ifdef SQR
-    let res = src[params.offset_src + src_idx] * src[params.offset_src + src_idx];
-#endif
-#ifdef SQRT
-    let res = sqrt(src[params.offset_src + src_idx]);
-#endif
-#ifdef SIN
-    let res_f32 = sin(f32(src[params.offset_src + src_idx]));
-    let res = TYPE(res_f32);
-#endif
-#ifdef COS
-    let res_f32 = cos(f32(src[params.offset_src + src_idx]));
-    let res = TYPE(res_f32);
-#endif
 
 #ifdef INPLACE
     src[params.offset_src + src_idx] = res;

ggml/src/ggml.c

@@ -899,7 +899,8 @@ static const struct ggml_type_traits type_traits[GGML_TYPE_COUNT] = {
 };
 
 const struct ggml_type_traits * ggml_get_type_traits(enum ggml_type type) {
-    GGML_ASSERT(type < GGML_TYPE_COUNT);
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
     return &type_traits[type];
 }
 
@@ -1265,27 +1266,33 @@ size_t ggml_nbytes_pad(const struct ggml_tensor * tensor) {
 }
 
 int64_t ggml_blck_size(enum ggml_type type) {
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
     return type_traits[type].blck_size;
 }
 
 size_t ggml_type_size(enum ggml_type type) {
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
     return type_traits[type].type_size;
 }
 
 size_t ggml_row_size(enum ggml_type type, int64_t ne) {
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
     assert(ne % ggml_blck_size(type) == 0);
     return ggml_type_size(type)*ne/ggml_blck_size(type);
 }
 
-double ggml_type_sizef(enum ggml_type type) {
-    return ((double)(type_traits[type].type_size))/type_traits[type].blck_size;
-}
-
 const char * ggml_type_name(enum ggml_type type) {
-    return type < GGML_TYPE_COUNT ? type_traits[type].type_name : "NONE";
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
+    return type_traits[type].type_name;
 }
 
 bool ggml_is_quantized(enum ggml_type type) {
+    assert(type >= 0);
+    assert(type < GGML_TYPE_COUNT);
     return type_traits[type].is_quantized;
 }
 
@@ -1629,11 +1636,23 @@ static struct ggml_object * ggml_new_object(struct ggml_context * ctx, enum ggml
     const size_t cur_end  = cur_offs + cur_size;
 
     // align to GGML_MEM_ALIGN
+    GGML_ASSERT(size <= SIZE_MAX - (GGML_MEM_ALIGN - 1));
     size_t size_needed = GGML_PAD(size, GGML_MEM_ALIGN);
 
     char * const mem_buffer = ctx->mem_buffer;
     struct ggml_object * const obj_new = (struct ggml_object *)(mem_buffer + cur_end);
 
+    // integer overflow checks
+    if (cur_end > SIZE_MAX - size_needed) {
+        GGML_LOG_WARN("%s: overflow detected in cur_end (%zu) + size_needed (%zu)\n", __func__, cur_end, size_needed);
+        return NULL;
+    }
+    if (cur_end + size_needed > SIZE_MAX - GGML_OBJECT_SIZE) {
+        GGML_LOG_WARN("%s: overflow detected in cur_end (%zu) + size_needed (%zu) + GGML_OBJECT_SIZE (%zu)\n", __func__,
+                cur_end, size_needed, (size_t) GGML_OBJECT_SIZE);
+        return NULL;
+    }
+
     if (cur_end + size_needed + GGML_OBJECT_SIZE > ctx->mem_size) {
         GGML_LOG_WARN("%s: not enough space in the context's memory pool (needed %zu, available %zu)\n",
                 __func__, cur_end + size_needed + GGML_OBJECT_SIZE, ctx->mem_size);
@@ -1702,6 +1721,8 @@ static struct ggml_tensor * ggml_new_tensor_impl(
         obj_alloc_size = data_size;
     }
 
+    GGML_ASSERT(GGML_TENSOR_SIZE <= SIZE_MAX - obj_alloc_size);
+
     struct ggml_object * const obj_new = ggml_new_object(ctx, GGML_OBJECT_TYPE_TENSOR, GGML_TENSOR_SIZE + obj_alloc_size);
     GGML_ASSERT(obj_new);
 

ggml/src/gguf.cpp

@@ -15,6 +15,9 @@
 #include <string>
 #include <vector>
 
+#define GGUF_MAX_STRING_LENGTH  (1024*1024*1024)
+#define GGUF_MAX_ARRAY_ELEMENTS (1024*1024*1024)
+
 template <typename T>
 struct type_to_gguf_type;
 
@@ -228,6 +231,26 @@ struct gguf_reader {
 
     template <typename T>
     bool read(std::vector<T> & dst, const size_t n) const {
+        if (n > GGUF_MAX_ARRAY_ELEMENTS) {
+            return false;
+        }
+        const uint64_t nbytes = nbytes_remain();
+        if constexpr (std::is_same<T, std::string>::value) {
+            // strings are prefixed with their length, so we need to account for that
+            if (n > SIZE_MAX / sizeof(uint64_t)) {
+                return false;
+            }
+            if (nbytes < n * sizeof(uint64_t)) {
+                return false;
+            }
+        } else {
+            if (n > SIZE_MAX / sizeof(T)) {
+                return false;
+            }
+            if (nbytes < n * sizeof(T)) {
+                return false;
+            }
+        }
         dst.resize(n);
         for (size_t i = 0; i < dst.size(); ++i) {
             if constexpr (std::is_same<T, bool>::value) {
@@ -277,13 +300,43 @@ struct gguf_reader {
         if (!read(size)) {
             return false;
         }
-        dst.resize(size);
+        if (size > GGUF_MAX_STRING_LENGTH) {
+            GGML_LOG_ERROR("%s: string length %" PRIu64 " exceeds maximum %" PRIu64 "\n", __func__, size, (uint64_t) GGUF_MAX_STRING_LENGTH);
+            return false;
+        }
+        const uint64_t nbytes = nbytes_remain();
+        if (size > nbytes) {
+            GGML_LOG_ERROR("%s: string length %" PRIu64 " exceeds remaining file size %" PRIu64 " bytes\n", __func__, size, nbytes);
+            return false;
+        }
+        dst.resize(static_cast<size_t>(size));
         return fread(dst.data(), 1, dst.length(), file) == dst.length();
     }
 
     bool read(void * dst, const size_t size) const {
         return fread(dst, 1, size, file) == size;
     }
+
+    // remaining bytes in the file
+    uint64_t nbytes_remain() const {
+        const long cur = ftell(file);
+        if (cur < 0) {
+            return 0;
+        }
+        if (fseek(file, 0, SEEK_END) != 0) {
+            fseek(file, cur, SEEK_SET);
+
+            return 0;
+        }
+        const long end = ftell(file);
+        if (end < 0) {
+            fseek(file, cur, SEEK_SET);
+
+            return 0;
+        }
+        fseek(file, cur, SEEK_SET);
+        return static_cast<uint64_t>(end - cur);
+    }
 };
 
 struct gguf_context * gguf_init_empty(void) {
@@ -568,8 +621,8 @@ struct gguf_context * gguf_init_from_file_impl(FILE * file, struct gguf_init_par
 
             // check that tensor type is within defined range
             if (info.t.type < 0 || info.t.type >= GGML_TYPE_COUNT) {
-                GGML_LOG_ERROR("%s: tensor '%s' has invalid ggml type %d (%s)\n",
-                    __func__, info.t.name, info.t.type, ggml_type_name(info.t.type));
+                GGML_LOG_ERROR("%s: tensor '%s' has invalid ggml type %d. should be in [0, %d)\n",
+                    __func__, info.t.name, info.t.type, GGML_TYPE_COUNT);
                 ok = false;
                 break;
             }
@@ -657,10 +710,34 @@ struct gguf_context * gguf_init_from_file_impl(FILE * file, struct gguf_init_par
         //   the ggml_tensor structs to the appropriate locations in the binary blob
 
         // compute the exact size needed for the new ggml_context
-        const size_t mem_size =
-            params.no_alloc ?
-            (n_tensors    )*ggml_tensor_overhead() :
-            (n_tensors + 1)*ggml_tensor_overhead() + ctx->size;
+        size_t mem_size = 0;
+        if (params.no_alloc) {
+            if (n_tensors != 0 && SIZE_MAX / n_tensors < ggml_tensor_overhead()) {
+                GGML_LOG_ERROR("%s: memory size overflow while allocating ggml context\n", __func__);
+                gguf_free(ctx);
+                return nullptr;
+            }
+
+            const size_t overhead = n_tensors * ggml_tensor_overhead();
+
+            mem_size = overhead;
+        } else {
+            if ((n_tensors + 1) != 0 && SIZE_MAX / (n_tensors + 1) < ggml_tensor_overhead()) {
+                GGML_LOG_ERROR("%s: memory size overflow while allocating ggml context\n", __func__);
+                gguf_free(ctx);
+                return nullptr;
+            }
+
+            const size_t overhead = (n_tensors + 1) * ggml_tensor_overhead();
+
+            if (SIZE_MAX - overhead < ctx->size) {
+                GGML_LOG_ERROR("%s: memory size overflow while allocating ggml context\n", __func__);
+                gguf_free(ctx);
+                return nullptr;
+            }
+
+            mem_size = overhead + ctx->size;
+        }
 
         struct ggml_init_params pdata = {
             /*mem_size   =*/ mem_size,

gguf-py/gguf/constants.py

@@ -473,7 +473,6 @@ class MODEL_ARCH(IntEnum):
     RND1             = auto()
     PANGU_EMBED      = auto()
     MISTRAL3         = auto()
-    PADDLEOCR        = auto()
     MIMO2            = auto()
     STEP35           = auto()
     LLAMA_EMBED      = auto()
@@ -915,7 +914,6 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
     MODEL_ARCH.RND1:             "rnd1",
     MODEL_ARCH.PANGU_EMBED:      "pangu-embedded",
     MODEL_ARCH.MISTRAL3:         "mistral3",
-    MODEL_ARCH.PADDLEOCR:        "paddleocr",
     MODEL_ARCH.MIMO2:            "mimo2",
     MODEL_ARCH.STEP35:           "step35",
     MODEL_ARCH.LLAMA_EMBED:      "llama-embed",
@@ -3188,20 +3186,6 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
         MODEL_TENSOR.FFN_DOWN,
         MODEL_TENSOR.FFN_UP,
     ],
-    MODEL_ARCH.PADDLEOCR: [
-        MODEL_TENSOR.TOKEN_EMBD,
-        MODEL_TENSOR.OUTPUT_NORM,
-        MODEL_TENSOR.OUTPUT,
-        MODEL_TENSOR.ATTN_NORM,
-        MODEL_TENSOR.ATTN_Q,
-        MODEL_TENSOR.ATTN_K,
-        MODEL_TENSOR.ATTN_V,
-        MODEL_TENSOR.ATTN_OUT,
-        MODEL_TENSOR.FFN_NORM,
-        MODEL_TENSOR.FFN_GATE,
-        MODEL_TENSOR.FFN_DOWN,
-        MODEL_TENSOR.FFN_UP,
-    ],
     MODEL_ARCH.FALCON_H1: [
         # Token embedding
         MODEL_TENSOR.TOKEN_EMBD,
@@ -3863,7 +3847,6 @@ class VisionProjectorType:
     VOXTRAL = "voxtral"
     LFM2 = "lfm2"
     KIMIVL = "kimivl"
-    PADDLEOCR = "paddleocr"
     KIMIK25 = "kimik25"
     LIGHTONOCR = "lightonocr"
     COGVLM = "cogvlm"

gguf-py/gguf/gguf_reader.py

@@ -175,6 +175,9 @@ class GGUFReader:
             if new_align.types != [GGUFValueType.UINT32]:
                 raise ValueError('Bad type for general.alignment field')
             self.alignment = new_align.parts[-1][0]
+            # Ensure alignment is a non-zero power of two
+            if self.alignment == 0 or (self.alignment & (self.alignment - 1)) != 0:
+                raise ValueError('Invalid alignment: must be a non-zero power of two')
         padding = offs % self.alignment
         if padding != 0:
             offs += self.alignment - padding
@@ -202,11 +205,11 @@ class GGUFReader:
 
     def _push_field(self, field: ReaderField, skip_sum: bool = False) -> int:
         if field.name in self.fields:
-            # TODO: add option to generate error on duplicate keys
-            # raise KeyError(f'Duplicate {field.name} already in list at offset {field.offset}')
+            # TODO: add option to make this a warning and accept duplicate keys like below
+            raise KeyError(f'Duplicate {field.name} already in list at offset {field.offset}')
 
-            logger.warning(f'Duplicate key {field.name} at offset {field.offset}')
-            self.fields[field.name + '_{}'.format(field.offset)] = field
+            # logger.warning(f'Duplicate key {field.name} at offset {field.offset}')
+            # self.fields[field.name + '_{}'.format(field.offset)] = field
         else:
             self.fields[field.name] = field
         return 0 if skip_sum else sum(int(part.nbytes) for part in field.parts)

gguf-py/gguf/gguf_writer.py

@@ -501,6 +501,8 @@ class GGUFWriter:
         self.add_uint32(Keys.General.QUANTIZATION_VERSION, quantization_version)
 
     def add_custom_alignment(self, alignment: int) -> None:
+        if alignment <= 0 or (alignment & (alignment - 1)) != 0:
+            raise ValueError('Invalid alignment: must be a non-zero power of two')
         self.data_alignment = alignment
         self.add_uint32(Keys.General.ALIGNMENT, alignment)
 

gguf-py/gguf/tensor_mapping.py

@@ -1325,7 +1325,6 @@ class TensorNameMap:
             "multi_modal_projector.linear_{bid}",
             "mm_projector.proj.linear_{bid}", # Kimi-K2.5
             "visual.merger.mlp.{bid}", # qwen2vl
-            "mlp_AR.linear_{bid}", # PaddleOCR-VL
             "merger.mlp.{bid}",
         ),
 
@@ -1575,7 +1574,6 @@ class TensorNameMap:
             "mm_projector.pre_norm", # Kimi-K2.5
             "pre_mm_projector_norm",
             "model.vision.linear_proj.norm1", # cogvlm
-            "mlp_AR.pre_norm", # PaddleOCR-VL
             "merger.ln_q",
         ),
 
@@ -1601,7 +1599,6 @@ class TensorNameMap:
 
         MODEL_TENSOR.V_RESMPL_ATTN_OUT: (
             "resampler.attn.out_proj",
-            "model.vision_model.head.attention.out_proj",
         ),
 
         MODEL_TENSOR.V_RESMPL_KV: (

src/CMakeLists.txt

@@ -110,7 +110,6 @@ add_library(llama
             models/openai-moe-iswa.cpp
             models/openelm.cpp
             models/orion.cpp
-            models/paddleocr.cpp
             models/pangu-embedded.cpp
             models/phi2.cpp
             models/phi3.cpp

src/llama-arch.cpp

@@ -121,7 +121,6 @@ static const std::map<llm_arch, const char *> LLM_ARCH_NAMES = {
     { LLM_ARCH_RND1,             "rnd1"             },
     { LLM_ARCH_PANGU_EMBED,      "pangu-embedded"   },
     { LLM_ARCH_MISTRAL3,         "mistral3"         },
-    { LLM_ARCH_PADDLEOCR,        "paddleocr"        },
     { LLM_ARCH_MIMO2,            "mimo2"            },
     { LLM_ARCH_STEP35,           "step35"           },
     { LLM_ARCH_LLAMA_EMBED,      "llama-embed"      },
@@ -740,7 +739,6 @@ static std::set<llm_tensor> llm_get_tensor_names(llm_arch arch) {
         case LLM_ARCH_INTERNLM2:
         case LLM_ARCH_GRANITE:
         case LLM_ARCH_ERNIE4_5:
-        case LLM_ARCH_PADDLEOCR:
         case LLM_ARCH_SMOLLM3:
         case LLM_ARCH_DREAM:
         case LLM_ARCH_LLADA:

src/llama-arch.h

@@ -125,7 +125,6 @@ enum llm_arch {
     LLM_ARCH_RND1,
     LLM_ARCH_PANGU_EMBED,
     LLM_ARCH_MISTRAL3,
-    LLM_ARCH_PADDLEOCR,
     LLM_ARCH_MIMO2,
     LLM_ARCH_STEP35,
     LLM_ARCH_LLAMA_EMBED,

src/llama-model.cpp

@@ -2244,11 +2244,7 @@ void llama_model::load_hparams(llama_model_loader & ml) {
             } break;
         case LLM_ARCH_ERNIE4_5:
         case LLM_ARCH_ERNIE4_5_MOE:
-        case LLM_ARCH_PADDLEOCR:
             {
-                // paddleocr need mrope_section
-                ml.get_key_or_arr(LLM_KV_ROPE_DIMENSION_SECTIONS, hparams.rope_sections, 4, false);
-
                 ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
                 if (arch == LLM_ARCH_ERNIE4_5_MOE) {
                     ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH,        hparams.n_ff_exp);
@@ -6635,7 +6631,6 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
                 } break;
             case LLM_ARCH_ERNIE4_5:
             case LLM_ARCH_ERNIE4_5_MOE:
-            case LLM_ARCH_PADDLEOCR:
                 {
                     tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
 
@@ -8714,10 +8709,6 @@ ggml_cgraph * llama_model::build_graph(const llm_graph_params & params) const {
             {
                 llm = std::make_unique<llm_build_ernie4_5_moe>(*this, params);
             } break;
-        case LLM_ARCH_PADDLEOCR:
-            {
-                llm = std::make_unique<llm_build_paddleocr>(*this, params);
-            } break;
         case LLM_ARCH_HUNYUAN_MOE:
             {
                 llm = std::make_unique<llm_build_hunyuan_moe>(*this, params);
@@ -9054,7 +9045,6 @@ llama_rope_type llama_model_rope_type(const llama_model * model) {
             return LLAMA_ROPE_TYPE_NEOX;
 
         case LLM_ARCH_QWEN2VL:
-        case LLM_ARCH_PADDLEOCR:
             return LLAMA_ROPE_TYPE_MROPE;
         case LLM_ARCH_QWEN3VL:
         case LLM_ARCH_QWEN3VLMOE:

src/llama-vocab.cpp

@@ -2470,7 +2470,6 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
                     || t.first == "<|calls|>"  // solar-open
                     || t.first == "<end_of_turn>"
                     || t.first == "<|endoftext|>"
-                    || t.first == "</s>"      // paddleocr
                     || t.first == "<|eom_id|>"
                     || t.first == "<EOT>"
                     || t.first == "_<EOT>"

src/models/models.h

@@ -190,10 +190,6 @@ struct llm_build_ernie4_5_moe : public llm_graph_context {
     llm_build_ernie4_5_moe(const llama_model & model, const llm_graph_params & params);
 };
 
-struct llm_build_paddleocr : public llm_graph_context {
-    llm_build_paddleocr(const llama_model & model, const llm_graph_params & params);
-};
-
 template <bool iswa>
 struct llm_build_exaone4 : public llm_graph_context {
     llm_build_exaone4(const llama_model & model, const llm_graph_params & params);

src/models/paddleocr.cpp

@@ -1,122 +0,0 @@
-#include "models.h"
-
-llm_build_paddleocr::llm_build_paddleocr(const llama_model & model, const llm_graph_params & params) :
-    llm_graph_context(params) {
-
-    // NOTE: same with qwen2vl.cpp, but bias tensors are optional
-
-    const int64_t n_embd_head = hparams.n_embd_head_v;
-
-    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
-    GGML_ASSERT(n_embd_head == hparams.n_rot);
-
-    ggml_tensor * cur;
-    ggml_tensor * inpL;
-
-    inpL = build_inp_embd(model.tok_embd);
-
-    int sections[4];
-    std::copy(std::begin(hparams.rope_sections), std::begin(hparams.rope_sections) + 4, sections);
-
-    // inp_pos - contains the positions
-    ggml_tensor * inp_pos = build_inp_pos();
-
-    auto * inp_attn = build_attn_inp_kv();
-
-    ggml_tensor * inp_out_ids = build_inp_out_ids();
-
-    for (int il = 0; il < n_layer; ++il) {
-        ggml_tensor * inpSA = inpL;
-
-        // norm
-        {
-            cur = build_norm(inpL, model.layers[il].attn_norm, NULL, LLM_NORM_RMS, il);
-            cb(cur, "attn_norm", il);
-        }
-        // self-attention
-        {
-            ggml_tensor * Qcur = build_lora_mm(model.layers[il].wq, cur);
-            cb(Qcur, "Qcur", il);
-            if (model.layers[il].bq) {
-                Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
-                cb(Qcur, "Qcur", il);
-            }
-            ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur);
-            cb(Kcur, "Kcur", il);
-            if (model.layers[il].bk) {
-                Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
-                cb(Kcur, "Kcur", il);
-            }
-            ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur);
-            cb(Vcur, "Vcur", il);
-            if (model.layers[il].bv) {
-                Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
-                cb(Vcur, "Vcur", il);
-            }
-            Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
-            Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
-            Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
-
-            Qcur = ggml_rope_multi(
-                    ctx0, Qcur, inp_pos, nullptr,
-                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,
-                    ext_factor, attn_factor, beta_fast, beta_slow
-                    );
-
-            Kcur = ggml_rope_multi(
-                    ctx0, Kcur, inp_pos, nullptr,
-                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,
-                    ext_factor, attn_factor, beta_fast, beta_slow
-                    );
-
-            cb(Qcur, "Qcur", il);
-            cb(Kcur, "Kcur", il);
-            cb(Vcur, "Vcur", il);
-
-            cur = build_attn(inp_attn,
-                    model.layers[il].wo, model.layers[il].bo,
-                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f/sqrtf(float(n_embd_head)), il);
-        }
-        if (il == n_layer - 1) {
-            // skip computing output for unused tokens
-            cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
-            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
-        }
-        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
-        cb(ffn_inp, "ffn_inp", il);
-
-        // feed-forward network
-        {
-            cur = build_norm(ffn_inp, model.layers[il].ffn_norm, NULL, LLM_NORM_RMS, il);
-            cb(cur, "ffn_norm", il);
-
-            cur = build_ffn(cur,
-                    model.layers[il].ffn_up, NULL, NULL,
-                    model.layers[il].ffn_gate, NULL, NULL,
-                    model.layers[il].ffn_down, NULL, NULL,
-                    NULL, LLM_FFN_SILU, LLM_FFN_PAR, il);
-            cb(cur, "ffn_out", il);
-        }
-        cur = ggml_add(ctx0, cur, ffn_inp);
-
-        cur = build_cvec(cur, il);
-        cb(cur, "l_out", il);
-
-        // input for next layer
-        inpL = cur;
-    }
-    cur = inpL;
-
-    cur = build_norm(cur, model.output_norm, NULL, LLM_NORM_RMS, -1);
-
-    cb(cur, "result_norm", -1);
-    res->t_embd = cur;
-
-    // lm_head
-    cur = build_lora_mm(model.output, cur);
-
-    cb(cur, "result_output", -1);
-    res->t_logits = cur;
-
-    ggml_build_forward_expand(gf, cur);
-}

tests/test-gguf.cpp

@@ -48,6 +48,7 @@ enum handcrafted_file_type {
     HANDCRAFTED_DATA_NOT_ENOUGH_DATA       =  10 + offset_has_data,
     HANDCRAFTED_DATA_BAD_ALIGN             =  15 + offset_has_data,
     HANDCRAFTED_DATA_INCONSISTENT_ALIGN    =  20 + offset_has_data,
+    HANDCRAFTED_DATA_MEM_SIZE_OVERFLOW     =  30 + offset_has_data,
     HANDCRAFTED_DATA_SUCCESS               = 800 + offset_has_data,
     HANDCRAFTED_DATA_CUSTOM_ALIGN          = 810 + offset_has_data,
 };
@@ -84,6 +85,7 @@ static std::string handcrafted_file_type_name(const enum handcrafted_file_type h
         case HANDCRAFTED_DATA_NOT_ENOUGH_DATA:       return "DATA_NOT_ENOUGH_DATA";
         case HANDCRAFTED_DATA_BAD_ALIGN:             return "DATA_BAD_ALIGN";
         case HANDCRAFTED_DATA_INCONSISTENT_ALIGN:    return "DATA_INCONSISTENT_ALIGN";
+        case HANDCRAFTED_DATA_MEM_SIZE_OVERFLOW:     return "DATA_MEM_SIZE_OVERFLOW";
         case HANDCRAFTED_DATA_SUCCESS:               return "DATA_SUCCESS";
         case HANDCRAFTED_DATA_CUSTOM_ALIGN:          return "DATA_CUSTOM_ALIGN";
     }
@@ -196,6 +198,13 @@ static FILE * get_handcrafted_file(const unsigned int seed, const enum handcraft
         tensor_configs = get_tensor_configs(rng);
     }
 
+    if (hft == HANDCRAFTED_DATA_MEM_SIZE_OVERFLOW) {
+        tensor_configs.resize(2);
+
+        tensor_configs[0] = { GGML_TYPE_I8, { 0x7FFFFFFFFFFFFFC0, 1, 1, 1 } };
+        tensor_configs[1] = { GGML_TYPE_I8, { 0x7FFFFFFFFFFFFFC0, 1, 1, 1 } };
+    }
+
     if (hft == HANDCRAFTED_HEADER_BAD_N_TENSORS) {
         const uint64_t n_tensors = -1;
         helper_write(file, n_tensors);
@@ -397,7 +406,8 @@ static FILE * get_handcrafted_file(const unsigned int seed, const enum handcraft
         for (uint32_t i = 1; i < n_dims; ++i) {
             ne *= shape[i];
         }
-        offset += GGML_PAD(ggml_row_size(type, ne), alignment);
+
+        offset += GGML_PAD(ggml_row_size(type, ne), (uint64_t) alignment);
     }
 
     while (ftell(file) % alignment != 0) {
@@ -411,6 +421,9 @@ static FILE * get_handcrafted_file(const unsigned int seed, const enum handcraft
         if (hft == HANDCRAFTED_DATA_NOT_ENOUGH_DATA) {
             nbytes -= 1;
         }
+        if (hft == HANDCRAFTED_DATA_MEM_SIZE_OVERFLOW) {
+            nbytes = 32;
+        }
         for (uint64_t i = 0; i < nbytes; ++i) {
             const uint8_t random_byte = i % 256;
             helper_write(file, random_byte);
@@ -704,6 +717,7 @@ static std::pair<int, int> test_handcrafted_file(const unsigned int seed) {
         HANDCRAFTED_DATA_NOT_ENOUGH_DATA,
         HANDCRAFTED_DATA_BAD_ALIGN,
         HANDCRAFTED_DATA_INCONSISTENT_ALIGN,
+        HANDCRAFTED_DATA_MEM_SIZE_OVERFLOW,
         HANDCRAFTED_DATA_SUCCESS,
         HANDCRAFTED_DATA_CUSTOM_ALIGN,
     };

tools/mtmd/CMakeLists.txt

@@ -24,7 +24,6 @@ add_library(mtmd
             models/llama4.cpp
             models/llava.cpp
             models/minicpmv.cpp
-            models/paddleocr.cpp
             models/pixtral.cpp
             models/qwen2vl.cpp
             models/qwen3vl.cpp

tools/mtmd/clip-impl.h

@@ -229,7 +229,6 @@ enum projector_type {
     PROJECTOR_TYPE_MUSIC_FLAMINGO,
     PROJECTOR_TYPE_LFM2,
     PROJECTOR_TYPE_KIMIVL,
-    PROJECTOR_TYPE_PADDLEOCR,
     PROJECTOR_TYPE_LIGHTONOCR,
     PROJECTOR_TYPE_COGVLM,
     PROJECTOR_TYPE_JANUS_PRO,
@@ -265,7 +264,6 @@ static std::map<projector_type, std::string> PROJECTOR_TYPE_NAMES = {
     { PROJECTOR_TYPE_MUSIC_FLAMINGO, "musicflamingo"},
     { PROJECTOR_TYPE_LFM2,      "lfm2"},
     { PROJECTOR_TYPE_KIMIVL,    "kimivl"},
-    { PROJECTOR_TYPE_PADDLEOCR, "paddleocr"},
     { PROJECTOR_TYPE_LIGHTONOCR,"lightonocr"},
     { PROJECTOR_TYPE_COGVLM,    "cogvlm"},
     { PROJECTOR_TYPE_JANUS_PRO, "janus_pro"},

tools/mtmd/clip.cpp

@@ -841,10 +841,6 @@ static ggml_cgraph * clip_image_build_graph(clip_ctx * ctx, const clip_image_f32
             {
                 builder = std::make_unique<clip_graph_kimivl>(ctx, img);
             } break;
-        case PROJECTOR_TYPE_PADDLEOCR:
-            {
-                builder = std::make_unique<clip_graph_paddleocr>(ctx, img);
-            } break;
         case PROJECTOR_TYPE_KIMIK25:
             {
                 builder = std::make_unique<clip_graph_kimik25>(ctx, img);
@@ -1260,14 +1256,6 @@ struct clip_model_loader {
                         hparams.audio_window_len   = 400;
                         hparams.audio_hop_len      = 160;
                     } break;
-                case PROJECTOR_TYPE_PADDLEOCR:
-                    {
-                        hparams.n_merge = 2;
-                        get_u32(KEY_IMAGE_MIN_PIXELS, hparams.image_min_pixels);
-                        get_u32(KEY_IMAGE_MAX_PIXELS, hparams.image_max_pixels);
-
-                        hparams.set_warmup_n_tokens(28*28); // avoid OOM on warmup
-                    } break;
                 case PROJECTOR_TYPE_LFM2A:
                     {
                         // audio preprocessing params
@@ -1716,7 +1704,6 @@ struct clip_model_loader {
                     model.mm_2_b = get_tensor(string_format(TN_LLAVA_PROJ, 2, "bias"));
                 } break;
             case PROJECTOR_TYPE_KIMIVL:
-            case PROJECTOR_TYPE_PADDLEOCR:
             case PROJECTOR_TYPE_KIMIK25:
                 {
                     model.mm_input_norm_w = get_tensor(TN_MM_INP_NORM);
@@ -3003,7 +2990,6 @@ bool clip_image_preprocess(struct clip_ctx * ctx, const clip_image_u8 * img, str
         case PROJECTOR_TYPE_QWEN25VL:
         case PROJECTOR_TYPE_QWEN3VL:
         case PROJECTOR_TYPE_GLM4V:
-        case PROJECTOR_TYPE_PADDLEOCR:
             {
                 GGML_ASSERT(params.image_min_pixels > 0 && params.image_max_pixels > 0);
                 clip_image_u8 resized;
@@ -3344,7 +3330,6 @@ int clip_n_output_tokens_x(const struct clip_ctx * ctx, struct clip_image_f32 *
         case PROJECTOR_TYPE_QWEN25VL:
         case PROJECTOR_TYPE_QWEN3VL:
         case PROJECTOR_TYPE_GLM4V:
-        case PROJECTOR_TYPE_PADDLEOCR:
         case PROJECTOR_TYPE_YOUTUVL:
             return (img->nx / params.patch_size) / 2;
         default:
@@ -3361,7 +3346,6 @@ int clip_n_output_tokens_y(const struct clip_ctx * ctx, struct clip_image_f32 *
         case PROJECTOR_TYPE_QWEN25VL:
         case PROJECTOR_TYPE_QWEN3VL:
         case PROJECTOR_TYPE_GLM4V:
-        case PROJECTOR_TYPE_PADDLEOCR:
         case PROJECTOR_TYPE_YOUTUVL:
             return (img->ny / params.patch_size) / 2;
         default:
@@ -3459,13 +3443,6 @@ int clip_n_output_tokens(const struct clip_ctx * ctx, struct clip_image_f32 * im
                 int y_patch = CLIP_ALIGN(img->ny, out_patch_size) / out_patch_size;
                 n_patches = x_patch * y_patch;
             } break;
-        case PROJECTOR_TYPE_PADDLEOCR:
-            {
-                // dynamic size
-                int n_merge = ctx->model.hparams.n_merge;
-                int stride = n_merge * n_merge;
-                n_patches = CLIP_ALIGN(n_patches, stride) / stride;
-            } break;
         case PROJECTOR_TYPE_PIXTRAL:
         case PROJECTOR_TYPE_LIGHTONOCR:
             {
@@ -3713,30 +3690,6 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
                     }
                 }
 
-                set_input_i32("positions", positions);
-            } break;
-        case PROJECTOR_TYPE_PADDLEOCR:
-            {
-                const int merge_ratio = hparams.n_merge;
-                const int pw = image_size_width  / patch_size;
-                const int ph = image_size_height / patch_size;
-                std::vector<int> positions(n_pos * 4);
-                int ptr = 0;
-                // NOTE: same as Qwen-VL, but x and y are swapped
-                for (int y = 0; y < ph; y += merge_ratio) {
-                    for (int dy = 0; dy < 2; dy++) {
-                        for (int x = 0; x < pw; x += merge_ratio) {
-                            for (int dx = 0; dx < 2; dx++) {
-                                positions[                  ptr] = y + dy;
-                                positions[    num_patches + ptr] = x + dx;
-                                positions[2 * num_patches + ptr] = y + dy;
-                                positions[3 * num_patches + ptr] = x + dx;
-                                ptr++;
-                            }
-                        }
-                    }
-                }
-
                 set_input_i32("positions", positions);
             } break;
         case PROJECTOR_TYPE_QWEN25VL:
@@ -4050,7 +4003,6 @@ int clip_n_mmproj_embd(const struct clip_ctx * ctx) {
             return ctx->model.mm_2_w->ne[1];
         case PROJECTOR_TYPE_LFM2:
         case PROJECTOR_TYPE_KIMIVL:
-        case PROJECTOR_TYPE_PADDLEOCR:
         case PROJECTOR_TYPE_KIMIK25:
             return ctx->model.mm_2_w->ne[1];
         case PROJECTOR_TYPE_COGVLM:

tools/mtmd/models/models.h

@@ -57,11 +57,6 @@ struct clip_graph_kimivl : clip_graph {
     ggml_cgraph * build() override;
 };
 
-struct clip_graph_paddleocr : clip_graph {
-    clip_graph_paddleocr(clip_ctx * ctx, const clip_image_f32 & img) : clip_graph(ctx, img) {}
-    ggml_cgraph * build() override;
-};
-
 struct clip_graph_cogvlm : clip_graph {
     clip_graph_cogvlm(clip_ctx * ctx, const clip_image_f32 & img) : clip_graph(ctx, img) {}
     ggml_cgraph * build() override;

tools/mtmd/models/paddleocr.cpp

@@ -1,52 +0,0 @@
-#include "models.h"
-
-ggml_cgraph * clip_graph_paddleocr::build() {
-    const int n_pos            = n_patches;
-    const int num_position_ids = n_pos * 4; // m-rope requires 4 dim per position
-
-    int mrope_sections[4] = {d_head/4, d_head/4, d_head/4, d_head/4};
-
-    ggml_tensor * positions = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, num_position_ids);
-    ggml_set_name(positions, "positions");
-    ggml_set_input(positions);
-
-    auto add_pos = [&](ggml_tensor * cur, const clip_layer &) {
-        return ggml_rope_multi(
-                    ctx0, cur, positions, nullptr,
-                    d_head/2, mrope_sections, GGML_ROPE_TYPE_VISION,
-                    32768, 10000, 1, 0, 1, 32, 1);
-    };
-
-    ggml_tensor * learned_pos_embd = resize_position_embeddings();
-    ggml_tensor * inp = build_inp();
-    ggml_tensor * cur = build_vit(
-                            inp, n_patches,
-                            NORM_TYPE_NORMAL,
-                            hparams.ffn_op,
-                            learned_pos_embd,
-                            add_pos);
-
-    cb(cur, "vit_out", -1);
-
-    {
-        // mlp_AR paddleocr projector
-        float proj_norm_eps = 1e-5;
-        cur = build_norm(cur,
-                    model.mm_input_norm_w, model.mm_input_norm_b,
-                    NORM_TYPE_NORMAL, proj_norm_eps, -1);
-
-        const int scale_factor = model.hparams.n_merge;
-        cur = build_patch_merge_permute(cur, scale_factor);
-        cur = build_ffn(cur,
-                    model.mm_1_w, model.mm_1_b,
-                    nullptr, nullptr,
-                    model.mm_2_w, model.mm_2_b,
-                    hparams.ffn_op, -1);
-        cb(cur, "mlp_out", -1);
-    }
-
-    // build the graph
-    ggml_build_forward_expand(gf, cur);
-
-    return gf;
-}

tools/mtmd/mtmd.cpp

@@ -325,10 +325,6 @@ struct mtmd_context {
             img_beg = "<|begin_of_image|>";
             img_end = "<|end_of_image|>";
 
-        } else if (proj == PROJECTOR_TYPE_PADDLEOCR) {
-            // <|IMAGE_START|> ... (image embeddings) ... <|IMAGE_END|>
-            img_beg = "<|IMAGE_START|>";
-            img_end = "<|IMAGE_END|>";
         }
     }
 
@@ -894,7 +890,6 @@ bool mtmd_decode_use_mrope(mtmd_context * ctx) {
         case PROJECTOR_TYPE_QWEN25VL:
         case PROJECTOR_TYPE_QWEN3VL:
         case PROJECTOR_TYPE_GLM4V:
-        case PROJECTOR_TYPE_PADDLEOCR:
             return true;
         default:
             return false;