Update README.md

README.md

@@ -98,6 +98,8 @@ print(generated_texts[0])
 
 ### Training Procedure
 
+See detailed blog on preference tuning VLLMs [here](https://huggingface.co/blog/dpo_vlm).
+
 ```bash
 accelerate launch  --config_file examples/accelerate_configs/multi_gpu.yaml \
   examples/scripts/dpo_vlm.py \
@@ -110,7 +112,7 @@ accelerate launch  --config_file examples/accelerate_configs/multi_gpu.yaml \
   --bf16 \
   --torch_dtype bfloat16 \
   --use_peft \
-  --lora_target_modules=all-linear exit
+  --lora_target_modules=all-linear
 ```
 
 ### Framework versions