README.md

---
license: mit
base_model:
- ibm-granite/granite-4.0-h-small
---


# Model Overview

- **Model Architecture:** Granite-4.0-h-small
  - **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:** FP8, Static
  - **Activation quantization:** FP8, Dynamic
- **Calibration Dataset:** [Pile](https://huggingface.co/datasets/mit-han-lab/pile-val-backup)

This model was built with deepseek-ai DeepSeek-R1-0528 model by applying [AMD-Quark](https://quark.docs.amd.com/latest/index.html) for MXFP4 quantization.

# Model Quantization

The model was quantized from [ibm-granite/granite-4.0-h-small](https://huggingface.co/ibm-granite/granite-4.0-h-small) 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-R1-0528-BF16](https://huggingface.co/unsloth/DeepSeek-R1-0528-BF16).

**Quantization scripts:**
```
cd Quark/examples/torch/language_modeling/llm_ptq/
exclude_layers="*router.* *lm_head"

python llm_ptq/quantize_quark.py \
                          --model_dir $MODEL_DIR \
                          --output_dir $OUT_DIR \
                          --quant_scheme w_fp8_a_fp8 \
                          --kv_cache_dtype fp8 \
                          --num_calib_data 128 \
                          --exclude_layers $exclude_layers \
                          --model_export hf_format \
                          --multi_gpu
```

# Deployment
### Use with SGLang

This model can be deployed efficiently using the [vllm](https://github.com/vllm-project/vllm) backend.

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