llama : change fallback type IQ4_NL -> Q4_0

ggml-ci
2026-02-12 14:03:20 +02:00 · 2024-07-16 10:00:57 +03:00
1 changed files with 4 additions and 4 deletions
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -18090,10 +18090,10 @@ static ggml_type llama_tensor_get_type(quantize_state_internal & qs, ggml_type n
            case GGML_TYPE_IQ1_M:
            case GGML_TYPE_Q2_K:
            case GGML_TYPE_Q3_K:
-            case GGML_TYPE_IQ4_XS: new_type = GGML_TYPE_IQ4_NL; break;
-            case GGML_TYPE_Q4_K:   new_type = GGML_TYPE_Q5_0;   break;
-            case GGML_TYPE_Q5_K:   new_type = GGML_TYPE_Q5_1;   break;
-            case GGML_TYPE_Q6_K:   new_type = GGML_TYPE_Q8_0;   break;
+            case GGML_TYPE_IQ4_XS: new_type = GGML_TYPE_Q4_0; break;
+            case GGML_TYPE_Q4_K:   new_type = GGML_TYPE_Q5_0; break;
+            case GGML_TYPE_Q5_K:   new_type = GGML_TYPE_Q5_1; break;
+            case GGML_TYPE_Q6_K:   new_type = GGML_TYPE_Q8_0; break;
            default: throw std::runtime_error("\nUnsupported tensor size encountered\n");
        }
        LLAMA_LOG_WARN(" - using fallback quantization %s\n", ggml_type_name(new_type));