Model Overview

Model Architecture: Qwen3_5MoeForConditionalGeneration
- Input: Text
- Output: Text
Supported Hardware Microarchitecture: AMD MI300 MI350/MI355
ROCm: 7.0.0
PyTorch: 2.9.1
Transformers: 5.3.0
Operating System(s): Linux
Inference Engine: SGLang/vLLM
Model Optimizer: AMD-Quark (v0.11.1)
- Weight quantization: OCP MXFP4, Static
- Activation quantization: OCP MXFP4, Dynamic

Model Quantization

The model was quantized from Qwen/Qwen3.5-397B-A17B-FP8 using AMD-Quark. The weights are quantized to MXFP4 and activations are quantized to MXFP4.

Quantization scripts:

import os
from quark.torch import LLMTemplate, ModelQuantizer
from quark.common.profiler import GlobalProfiler

# Register qwen3_5_moe template
qwen3_5_moe_template = LLMTemplate(
    model_type="qwen3_5_moe",
    kv_layers_name=["*k_proj", "*v_proj"],
    q_layer_name="*q_proj"
)
LLMTemplate.register_template(qwen3_5_moe_template)

# Configuration
ckpt_path = "Qwen/Qwen3.5-397B-A17B-FP8"
output_dir = "amd/Qwen3.5-397B-A17B-MXFP4"
quant_scheme = "mxfp4"
exclude_layers = ["lm_head", "model.visual.*", "mtp.*", "*mlp.gate", "*shared_expert_gate*", "*.linear_attn.*", "*.self_attn.*", "*.shared_expert.*"]

# Get quant config from template
template = LLMTemplate.get("qwen3_5_moe")
quant_config = template.get_config(scheme=quant_scheme, exclude_layers=exclude_layers)

# Quantize with File-to-file mode
profiler = GlobalProfiler(output_path=os.path.join(output_dir, "quark_profile.yaml"))
quantizer = ModelQuantizer(quant_config)
quantizer.direct_quantize_checkpoint(
    pretrained_model_path=ckpt_path,
    save_path=output_dir,
)

For further details or issues, please refer to the AMD-Quark documentation or contact the respective developers.

Evaluation

The model was evaluated on gsm8k benchmarks using the vllm framework.

Accuracy

Benchmark	Qwen/Qwen3.5-397B-A17B-FP8	amd/Qwen3.5-397B-A17B-MXFP4(this model)	Recovery
gsm8k (flexible-extract)	95.38	94.24	98.80%

Reproduction

The GSM8K results were obtained using the vLLM framework, based on the Docker image rocm/vllm-dev:nightly_main_20260211, and vLLM is installed inside the container.

Evaluating model in a new terminal

lm_eval \
  --model vllm \
  --model_args pretrained=$MODEL,tensor_parallel_size=1,max_model_len=262144,gpu_memory_utilization=0.90,max_gen_toks=2048,trust_remote_code=True,reasoning_parser=qwen3 \
  --tasks gsm8k  --num_fewshot 5 \
  --batch_size auto