Harahan/MeanFlowNFT

Bringing Forward-Process RL to Average-Velocity Generators

Yushi Huang^1,2* · Xiangxin Zhou¹*^✉ · Jun Zhang² · Liefeng Bo¹ · Tianyu Pang^1,✉

¹Tencent Hunyuan    ²The Hong Kong University of Science and Technology

* Equal contribution    ^✉Corresponding authors

MeanFlowNFT optimizes reward in induced instantaneous-velocity space while retaining the average-velocity parameterization and native any-step MeanFlow sampler. Training uses only forward-noised clean samples, without reverse-trajectory likelihood estimation.

🍭 Overview

Optimize in instantaneous-velocity space while preserving MeanFlow sampling.

One practical MeanFlowNFT update.

📦 Contents

This repository contains the three ordered adapter components required to reconstruct the final policy:

.
├── anyflow_pretrain/
│   └── generator_ema.pt
├── anyflow_onpolicy/
│   └── generator_ema.pt
└── meanflownft/
    └── generator_ema.pt

The loading contract is strict:

1. Merge anyflow_pretrain/generator_ema.pt into the base transformer.
2. Merge anyflow_onpolicy/generator_ema.pt on top of that result.
3. Keep meanflownft/generator_ema.pt as the active final adapter.
4. Load the complete delta_embedder state from the final checkpoint.

The adapter checkpoints use LoRA rank 32, alpha 64, and attention projections matching the released code. They also contain the trainable flow-map delta_embedder, so they are not conventional LoRA-only files. SHA256SUMS records the checksum of every checkpoint.

📊 Results

MeanFlowNFT is best on 6 of 8 image metrics among the evaluated few-step models, while 4-step Wan2.1 reaches 84.33 VBench, surpassing 50-step LongCat-Video RL. See the project page for full comparisons, scaling curves, and qualitative results.

🚀 Usage

Use the MeanFlowNFT codebase, whose inference loader preserves the ordered adapter composition and full delta_embedder state.

git clone https://github.com/Harahan/MeanFlowNFT.git
cd MeanFlowNFT
pip install -r requirements.txt
pip install -e .

export HF_REPO_ID=Harahan/MeanFlowNFT
hf download "${HF_REPO_ID}" --local-dir ./checkpoints/meanflownft

hf auth login
hf download stabilityai/stable-diffusion-3.5-medium \
    --local-dir ./models/stable-diffusion-3.5-medium

Run the final 4-step policy:

python inference.py configs/inference/sd35m_meanflow_nft.yaml \
    --override \
    pretrained_path=./models/stable-diffusion-3.5-medium \
    num_stages=3 \
    stage1_lora_path=./checkpoints/meanflownft/anyflow_pretrain/generator_ema.pt \
    stage2_lora_path=./checkpoints/meanflownft/anyflow_onpolicy/generator_ema.pt \
    stage3_lora_path=./checkpoints/meanflownft/meanflownft/generator_ema.pt \
    num_steps=4 \
    eval_reward=false \
    --prompt "a cinematic photo of a red panda"

The first two adapters are merged in order. The final MeanFlowNFT adapter stays active and supplies delta_embedder, matching training-time evaluation. The released inference config generates 1024×1024 images with CFG-free flow-map sampling.

Do not load only the final checkpoint and do not use a generic LoRA loader that discards non-LoRA tensors. Either mistake drops part of the trained policy.

These .pt files use PyTorch serialization. Load checkpoints only from a trusted source.

📄 Citation

@misc{huang2026meanflownftbringingforwardprocessrl,
  title  = {MeanFlowNFT: Bringing Forward-Process RL to Average-Velocity Generators},
  author = {Huang, Yushi and Zhou, Xiangxin and Zhang, Jun and Bo, Liefeng and Pang, Tianyu},
  year={2026},
  eprint={2607.15273},
  archivePrefix={arXiv},
  primaryClass={cs.CV},
  url={https://arxiv.org/abs/2607.15273}, 
}

🙌 Acknowledgements

This work builds on MeanFlow, AnyFlow, DiffusionNFT, Diffusers, Transformers, and PEFT.

⚖️ License

The repository metadata and accompanying code are released under Apache 2.0. The base model and all derivative weights remain subject to the base model's license and usage terms; review them before use or redistribution.