README.md

---
license: mit
base_model:
- deepseek-ai/DeepSeek-R1
---


# Model Overview

- **Model Architecture:** DeepSeek-R1
  - **Input:** Text
  - **Output:** Text
- **Supported Hardware Microarchitecture:** AMD MI350/MI355
- **ROCm**: 7.0
- **Operating System(s):** Linux
- **Inference Engine:** [SGLang](https://docs.sglang.ai/)
- **Model Optimizer:** [AMD-Quark](https://quark.docs.amd.com/latest/index.html)
  - **Weight quantization:** OCP MXFP4, Static
  - **Activation quantization:** OCP MXFP4, Dynamic
- **Calibration Dataset:** [Pile](https://huggingface.co/datasets/mit-han-lab/pile-val-backup)

This model was built with deepseek-ai DeepSeek-R1 model by applying [AMD-Quark](https://quark.docs.amd.com/latest/index.html) for MXFP4 quantization.

# Model Quantization

The model was quantized from [deepseek-ai/DeepSeek-R1](https://huggingface.co/deepseek-ai/DeepSeek-R1) using [AMD-Quark](https://quark.docs.amd.com/latest/index.html). Both weights and activations were quantized to MXFP4 format, and the AutoSmoothQuant algorithm was applied to enhance accuracy. 

**Preprocessing requirement:**

Before executing the quantization script below, the original FP8 model must first be dequantized to BFloat16.
You can either perform the dequantization manually using this [conversion script](https://github.com/deepseek-ai/DeepSeek-V3/blob/main/inference/fp8_cast_bf16.py), or use the pre-converted BFloat16 model available at [unsloth/DeepSeek-R1-BF16](https://huggingface.co/unsloth/DeepSeek-R1-BF16).

**Quantization scripts:**
```
cd Quark/examples/torch/language_modeling/llm_ptq/
exclude_layers="*self_attn* *mlp.gate.* *lm_head"
python3 quantize_quark.py --model_dir $MODEL_DIR \
                          --quant_scheme w_mxfp4_a_mxfp4 \
                          --group_size 32 \
                          --num_calib_data 128 \
                          --exclude_layers $exclude_layers \
                          --multi_gpu \
                          --quant_algo autosmoothquant \
                          --model_export hf_format \
                          --output_dir amd/DeepSeek-R1-MXFP4-ASQ
```

# Deployment
### Use with SGLang

This model can be deployed efficiently using the [SGLang](https://docs.sglang.ai/) backend.

## Evaluation

The model was evaluated on reasoning tasks including AIME24, MMLU_COT, and GSM8K via [forked lm-evaluation-harness](https://github.com/BowenBao/lm-evaluation-harness/tree/cot) .

### Accuracy

<table>
  <tr>
   <td><strong>Benchmark</strong>
   </td>
   <td><strong>DeepSeek-R1 </strong>
   </td>
   <td><strong>DeepSeek-R1-MXFP4-ASQ(this model)</strong>
   </td>
   <td><strong>Recovery</strong>
   </td>
  </tr>
  <tr>
   <td>AIME24 
   </td>
   <td>78.0
   </td>
   <td>76.0
   </td>
   <td>97.44%
   </td>
  </tr>
  <tr>
   <td>MMLU_COT
   </td>
   <td>79.90
   </td>
   <td>79.65
   </td>
   <td>99.69%
   </td>
  </tr>
  <tr>
   <td>GSM8K
   </td>
   <td>95.81
   </td>
   <td>95.42 
   </td>
   <td>99.59%
   </td>
  </tr>
</table>


### Reproduction

The results of AIME24 and MMLU_COT were obtained using [SGLang](https://docs.sglang.ai/) while result of GSM8K was obtained using [vLLM](https://docs.vllm.ai/en/latest/). All the evaluations were conducted via forked [lm-evaluation-harness](https://github.com/BowenBao/lm-evaluation-harness/tree/cot).

### AIME24
```
# Launching server
python3 -m sglang.launch_server \
    --model /data/DeepSeek-R1-WMXFP4-AMXFP4-Scale-UINT8-Attn-MoE-Quant/ \
    --tp 8  \
    --trust-remote-code  \
    --n-share-experts-fusion 8 \
    --disable-radix-cache

# Evaluating
lm_eval --model local-completions \
    --model_args model=amd/DeepSeek-R1-MXFP4-ASQ,base_url=http://localhost:30000/v1/completions,num_concurrent=999999,timeout=999999,tokenized_requests=False,max_length=32000,temperature=0.6,top_p=0.95 \
    --tasks aime24 \
    --num_fewshot 0 \
    --gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,max_tokens=32000" \
    --batch_size auto \
    --log_samples \
    --output_path output_data/aime24 2>&1 | tee logs/aime24.log
```

### MMLU_COT
```
# Launching server
python3 -m sglang.launch_server \
    --model amd/DeepSeek-R1-MXFP4-ASQ \
    --tp 8 \
    --trust-remote-code \
    --chunked-prefill-size 32768 \
    --mem-fraction-static 0.83

# Evaluating
lm_eval --model local-completions \
    --model_args model=amd/DeepSeek-R1-MXFP4-ASQ,base_url=http://localhost:30000/v1/completions,num_concurrent=999999,timeout=999999,tokenized_requests=False,max_length=32000,temperature=0.6,top_p=0.95 \
    --tasks mmlu_cot \
    --num_fewshot 0 \
    --gen_kwargs "do_sample=True,temperature=0.6,top_p=0.95,max_tokens=32000" \
    --batch_size auto \
    --log_samples \
    --output_path output_data/mmmlu_cot 2>&1 | tee logs/mmmlu_cot.log
```


### GSM8K
```
lm_eval --model local-completions \
    --model_args model=amd/DeepSeek-R1-MXFP4-ASQ,base_url=http://localhost:30000/v1/completions,num_concurrent=999999,timeout=999999,tokenized_requests=False,max_length=8096 \
    --tasks gsm8k \
    --num_fewshot 5 \
    --batch_size auto \
    --log_samples \
    --output_path output_data/gsm8k 2>&1 | tee logs/gsm8k.log
```


# License
Modifications Copyright(c) 2025 Advanced Micro Devices, Inc. All rights reserved.