update readme and documentation

Signed-off-by: Suguna Velury <[email protected]>

Author: sugunav14

examples/llm_qat/README.md

@@ -350,7 +350,7 @@ After performing QLoRA training the final checkpoint can be exported for deploym
 ```sh
 python export.py \
    --pyt_ckpt_path llama3-fp4-qlora \
-   --export_dir llama3-fp4-qlora-hf \
+   --export_path llama3-fp4-qlora-hf \
 
 ```
 

modelopt/torch/export/quant_utils.py

@@ -827,6 +827,7 @@ def postprocess_state_dict(
         state_dict: The full model state_dict.
         maxbound: The maximum bound value for the output quantizer.
         quantization: The KV cache quantization format.
+        is_modelopt_qlora: Whether the model is a modelopt-trained QLoRA model.
 
     Returns:
         The filtered state_dict without unnecessary keys like '_amax' and non KV cache output quantizers.