[Feature] Support NVFP4 MoE

[Echo-Nie]

Motivation

This pr supports modelopt format NVFP4 inference (currently, only Qwen/Ernie) by introducing Flashinfer as a backend. It requires GPU with sm>=100 and Flashinfer installtion.

Modifications

With paddle compatible api, this pr introdudaces Flashinfer as a backend. There may be coexistence issues with some third-party including pytorch code (e.g. xgrammar, triton). , Currenly we cannot use them at same time and we are working on resolving this.

Usage or Command

New Environment Variables

• FD_FLASHINFER_MOE_BACKEND : FP4 MoE backend, could be flashinfer-cutlass, flashinfer-trtllm or None (default is None, will use flashinfer-cutass). Currently we only support flashinfer-cutlass.

• FD_NVFP4_GEMM_BACKEND: FP4 dense GEMM backend, could be flashinfer-cutlass, flashinfer-trtllm, flashinfer-cudnn or None (default is None, will use flashinfer-cutlass). Currently we only support flashinfer-cutlass.

• PADDLE_COMPATIBLE_API: This is an environment variable for Flashinfer with Paddle, set it to true to use paddle compatible api, default is false.

Start the Server

python -m fastdeploy.entrypoints.openai.api_server \
    --model nv-community/Qwen3-30B-A3B-FP4 \
    --port 8180 \
    --metrics-port 8181 \
    --engine-worker-queue-port 8182 \
    --cache-queue-port 8183 \
    --tensor-parallel-size 1 \
    --max-model-len  32768 \
    --max-num-seqs 128

Performance Benchmark Command

Reference: https://github.com/PaddlePaddle/FastDeploy/tree/develop/benchmarks

python benchmark_serving.py \
  --backend openai-chat \
  --model nv-community/Qwen3-30B-A3B-FP4 \
  --endpoint /v1/chat/completions \
  --host 0.0.0.0 \
  --port 8180 \
  --dataset-name EBChat \
  --dataset-path ./data/filtered_sharedgpt_2000_input_1136_output_200_fd.json \
  --hyperparameter-path ./yaml/request_yaml/qwen25-vl-32kyaml \
  --percentile-metrics ttft,tpot,itl,e2el,s_ttft,s_itl,s_e2el,s_decode,input_len,s_input_len,output_len \
  --metric-percentiles 80,95,99,99.9,99.95,99.99 \
  --num-prompts 1000 \
  --max-concurrency 64 \
  --save-result

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!

[codecov-commenter]

Codecov Report

❌ Patch coverage is 15.51155% with 256 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@fe5ba4b). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...deploy/model_executor/layers/quantization/nvfp4.py	12.79%	224 Missing and 1 partial ⚠️
fastdeploy/model_executor/layers/moe/moe.py	20.00%	19 Missing and 1 partial ⚠️
fastdeploy/flashinfer.py	63.63%	2 Missing and 2 partials ⚠️
...loy/model_executor/layers/quantization/__init__.py	20.00%	4 Missing ⚠️
fastdeploy/model_executor/utils.py	25.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #6003   +/-   ##
==========================================
  Coverage           ?   67.28%           
==========================================
  Files              ?      355           
  Lines              ?    46098           
  Branches           ?     7111           
==========================================
  Hits               ?    31015           
  Misses             ?    12829           
  Partials           ?     2254           

Flag	Coverage Δ
GPU	67.27% <15.51%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Echo-Nie]

/re-run all-failed

[Echo-Nie]

/re-run base_tests

[Echo-Nie]

/re-run all-failed

[Echo-Nie]

/re-run all-failed

[Echo-Nie]

/re-run all-failed