Update README.md

README.md

@@ -25,7 +25,8 @@ python3 quantize_quark.py \
         --quant_scheme w_fp8_a_fp8 \
         --kv_cache_dtype fp8 \
         --num_calib_data 128  \
-        --model_export quark_safetensors
+        --model_export quark_safetensors \
+        --no_weight_matrix_merge
 
 # If model size is too large for single GPU, please use multi GPU instead.
 python3 quantize_quark.py \ 
@@ -35,6 +36,7 @@ python3 quantize_quark.py \
         --kv_cache_dtype fp8 \
         --num_calib_data 128  \
         --model_export quark_safetensors \
+        --no_weight_matrix_merge \
         --multi_gpu
 ```
 ## Deployment
@@ -57,9 +59,9 @@ The quantization evaluation results are conducted in pseudo-quantization mode, w
   <tr>
    <td>Perplexity-wikitext2
    </td>
-   <td>1.8561
+   <td>1.8562
    </td>
-   <td>1.8951
+   <td>1.8927
    </td>
   </tr>