Update README.md

README.md

@@ -16,8 +16,8 @@ base_model:
 - **Operating System(s):** Linux
 - **Inference Engine:** [vLLM](https://docs.vllm.ai/en/latest/)
 - **Model Optimizer:** [AMD-Quark](https://quark.docs.amd.com/latest/index.html)
-  - **Weight quantization:** OCP MXFP4, Static, MOE-only
-  - **Activation quantization:** OCP MXFP4, Dynamic, MOE-only
+  - **Weight quantization:** MOE-only, OCP MXFP4, Static 
+  - **Activation quantization:** MOE-only, OCP MXFP4, Dynamic
 - **Calibration Dataset:** [Pile](https://huggingface.co/datasets/mit-han-lab/pile-val-backup)
 
 This model was built with Kimi-K2-Thinking model by applying [AMD-Quark](https://quark.docs.amd.com/latest/index.html) for MXFP4 quantization.
@@ -75,13 +75,10 @@ The model was evaluated on GSM8K benchmarks.
 </table>
 
 ### Reproduction
-The GSM8K results were obtained using the [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness.git) framework, based on the Docker image `rocm/vllm-dev:base`, with vLLM and lm-eval compiled and installed from source inside the container.
 
-#### Commit Hash:
-- **vLLM:** [cbbae38f9368b6c35d9b9295bf4ceee1e6452750](https://github.com/vllm-project/vllm/commit/cbbae38f9368b6c35d9b9295bf4ceee1e6452750)
-- **Quark:** [1742e8c40f8b90c4ecb2b086788160e919986399](https://gitenterprise.xilinx.com/AMDNeuralOpt/Quark/commit/1742e8c40f8b90c4ecb2b086788160e919986399)
-- **lm-evaluation-harness:** [4b74ec1268267ea2ea83893400d7013df30507af](https://github.com/EleutherAI/lm-evaluation-harness/commit/4b74ec1268267ea2ea83893400d7013df30507af)
-  - tips: Remove the 'hf' and 'vllm' optional dependency lines from pyproject.toml in the lm-evaluation-harness repository to prevent installation errors.
+GSM8K evaluation was conducted using the lm-evaluation-harness framework within the Docker image `rocm/vllm-dev:base`. The inference backend was vLLM, built from source at commit [cbbae38f9368b6c35d9b9295bf4ceee1e6452750](https://github.com/vllm-project/vllm/commit/cbbae38f9368b6c35d9b9295bf4ceee1e6452750).
+To ensure reproducibility and to avoid environment and dependency inconsistencies, we relied on the Docker-prebuilt `lm-evaluation-harness`, corresponding to commit [4b74ec1268267ea2ea83893400d7013df30507af](https://github.com/EleutherAI/lm-evaluation-harness/commit/4b74ec1268267ea2ea83893400d7013df30507af).
+
 
 #### Launching server
 ```