Update README.md

README.md

@@ -163,7 +163,6 @@ lm-eval --model hf --model_args pretrained=OPEA/DeepSeek-R1-Distill-Qwen-32B-int
 |           Metric           |   BF16   |      INT4      |
 | :------------------------: | :----------------------: | :---------------: |
 | avg                  | 0.6647 |  0.6639|
-|----------------------|--------|--------|
 | leaderboard_mmlu_pro | -      | -      |
 | mmlu                 | 0.7964 | 0.7928 |
 | lambada_openai       | 0.6649 | 0.6718 |
@@ -183,7 +182,17 @@ lm-eval --model hf --model_args pretrained=OPEA/DeepSeek-R1-Distill-Qwen-32B-int
 ### Generate the model
 
 ~~~bash
+Here is the sample command to generate the model.
 
+```bash
+auto-round  \
+--model  \
+--device 0 \
+--bits 4 \
+--iter 200 \
+--disable_eval \
+--format 'auto_gptq,auto_round' \
+--output_dir "./tmp_autoround" 
 
 ~~~