Change imatrix format to gguf

app.py

@@ -152,7 +152,6 @@ class HuggingFaceModelProcessor:
             "-f", train_data_path,
             "-ngl", "99",
             "--output-frequency", "10",
-            "--output-format", "dat",
             "-o", quant_config.imatrix_file,
         ]
 
@@ -417,7 +416,7 @@ llama-server --hf-repo "{processing_config.new_repo_id}" --hf-file "{processing_
         if quant_config.use_imatrix and os.path.isfile(quant_config.imatrix_file):
             try:
                 print(f"Uploading imatrix.dat: {os.path.abspath(quant_config.imatrix_file)}")
-                self._upload_file(processing_config, quant_config.imatrix_file, f"{processing_config.model_name}-imatrix.dat")
+                self._upload_file(processing_config, quant_config.imatrix_file, f"{processing_config.model_name}-imatrix.gguf")
             except Exception as e:
                 raise GGUFConverterError(f"Error uploading imatrix.dat: {e}")
 
@@ -783,7 +782,7 @@ class GGUFConverterUI:
                 )
 
                 quant_config.fp16_model = f"{outdir}/{model_name}-fp16.gguf"
-                quant_config.imatrix_file = f"{outdir}/{model_name}-imatrix.dat"
+                quant_config.imatrix_file = f"{outdir}/{model_name}-imatrix.gguf"
                 quant_config.quantized_gguf = f"{outdir}/{gguf_name}"
 
                 processing_config = ModelProcessingConfig(