This GGUF file is a direct conversion of Wan-AI/Wan2.2-T2V-A14B Since this is a quantized model, all original licensing terms and usage restrictions remain in effect. Usage The model can be used with the ComfyUI custom node ComfyUI-GGUF by city96 Place model files in ComfyUI/models/unet see the GitHub readme for further installation instructions.
Open-weight model · Text to video
Wan2.2-Fun-Reward-LoRAs
by Alibaba-PAI alibaba-pai/Wan2.2-Fun-Reward-LoRAs
We explore the Reward Backpropagation technique 1 2 to optimized the generated videos by Wan2.2-Fun for better alignment with human preferences. We provide the following pre-trained models (i.e. LoRAs) along with the training script.
Parameters—
Context—
Weights3.4 GB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads23.9k
Model Card
By Alibaba-PAI, published under apache-2.0, revision 98ccd8617f31.
We explore the Reward Backpropagation technique 1 2 to optimized the generated videos by Wan2.2-Fun for better alignment with human preferences. We provide the following pre-trained models (i.e. LoRAs) along with the training script. You can use these LoRAs to enhance the corresponding base model as a plug-in or train your own reward LoRA. For more details, please refer to our GitHub repo. A panda eats bamboo while a monkey swings from branch to branch A dog runs through a field while a cat climbs a tree A penguin waddles on the ice, a camel treks by Pig with wings flying above a diamond mountain Set lorapath along with loraweight for the low noise reward LoRA, while specifying lorahighpath…
Read Alibaba-PAI's full model card
Introduction
We explore the Reward Backpropagation technique 1 2 to optimized the generated videos by Wan2.2-Fun for better alignment with human preferences. We provide the following pre-trained models (i.e. LoRAs) along with the training script. You can use these LoRAs to enhance the corresponding base model as a plug-in or train your own reward LoRA.
For more details, please refer to our GitHub repo.
| Name | Base Model | Reward Model | Hugging Face | Description |
|---|---|---|---|---|
| Wan2.2-Fun-A14B-InP-high-noise-HPS2.1.safetensors | Wan2.2-Fun-A14B-InP (high noise) | HPS v2.1 | Link | Official HPS v2.1 reward LoRA (rank=128 and network_alpha=64) for Wan2.2-Fun-A14B-InP (high noise). It is trained with a batch size of 8 for 5,000 steps. |
| Wan2.2-Fun-A14B-InP-low-noise-HPS2.1.safetensors | Wan2.2-Fun-A14B-InP (low noise) | MPS | Link | Official HPS v2.1 reward LoRA (rank=128 and network_alpha=64) for Wan2.2-Fun-A14B-InP (low noise). It is trained with a batch size of 8 for 2,700 steps. |
| Wan2.2-Fun-A14B-InP-high-noise-MPS.safetensors | Wan2.2-Fun-A14B-InP (high noise) | HPS v2.1 | Link | Official MPS reward LoRA (rank=128 and network_alpha=64) for Wan2.2-Fun-A14B-InP (high noise). It is trained with a batch size of 8 for 5,000 steps. |
| Wan2.2-Fun-A14B-InP-low-noise-MPS.safetensors | Wan2.2-Fun-A14B-InP (low noise) | MPS | Link | Official MPS reward LoRA (rank=128 and network_alpha=64) for Wan2.2-Fun-A14B-InP (low noise). It is trained with a batch size of 8 for 4,500 steps. |
[!NOTE] We found that, MPS reward LoRA for the low-noise model converges significantly more slowly than on the other models, and may not deliver satisfactory results. Therefore, for the low-noise model, we recommend using HPSv2.1 reward LoRA.
Demo
Wan2.2-Fun-A14B-InP
| Prompt | Wan2.2-Fun-A14B-InP | Wan2.2-Fun-A14B-InP high + low HPSv2.1 Reward LoRA |
Wan2.2-Fun-A14B-InP high MPS + low HPSv2.1 Reward LoRA |
|---|---|---|---|
A panda eats bamboo while a monkey swings from branch to branch
ExpandedIn a lush green forest, a panda sits comfortably against a tree, leisurely munching on bamboo stalks. Nearby, a lively monkey swings energetically from branch to branch, its tail curling around the limbs. Sunlight filters through the canopy, casting dappled shadows on the forest floor. |
|||
A dog runs through a field while a cat climbs a tree
ExpandedIn a sunlit, expansive green field surrounded by tall trees, a playful golden retriever sprints energetically across the grass, its fur gleaming in the afternoon sun. Nearby, a nimble tabby cat gracefully climbs a sturdy tree, its claws gripping the bark effortlessly. The sky is clear blue with occasional birds flying. |
|||
A penguin waddles on the ice, a camel treks by
ExpandedA small penguin waddles slowly across a vast, icy surface under a clear blue sky. The penguin's short, flipper-like wings sway at its sides as it moves. Nearby, a camel treks steadily, its long legs navigating the snowy terrain with ease. The camel's fur is thick, providing warmth in the cold environment. |
|||
Pig with wings flying above a diamond mountain
ExpandedA whimsical pig, complete with delicate feathered wings, soars gracefully above a shimmering diamond mountain. The pig's pink skin glistens in the sunlight as it flaps its wings. The mountain below sparkles with countless facets, reflecting brilliant rays of light into the clear blue sky. |
[!NOTE] The above test prompts are from T2V-CompBench and expanded into detailed prompts by Llama-3.3. Videos are generated with HPSv2.1 Reward LoRA weight 0.5 and MPS Reward LoRA weight 0.5.
Quick Start
Set lora_path along with lora_weight for the low noise reward LoRA, while specifying lora_high_path and lora_high_weight for high noise reward LoRA in examples/wan2.2_fun/predict_t2v.py.
Training
Please refer to README_TRAIN_REWARD.md
Limitations
- We observe after training to a certain extent, the reward continues to increase, but the quality of the generated videos does not further improve. The model trickly learns some shortcuts (by adding artifacts in the background, i.e., adversarial patches) to increase the reward.
- Currently, there is still a lack of suitable preference models for video generation. Directly using image preference models cannot evaluate preferences along the temporal dimension (such as dynamism and consistency). Further more, We find using image preference models leads to a decrease in the dynamism of generated videos. Although this can be mitigated by computing the reward using only the first frame of the decoded video, the impact still persists.
Reference
- Clark, Kevin, et al. "Directly fine-tuning diffusion models on differentiable rewards.". In ICLR 2024.
- Prabhudesai, Mihir, et al. "Aligning text-to-image diffusion models with reward backpropagation." arXiv preprint arXiv:2310.03739 (2023).
Identity and Version
- Repository
- alibaba-pai/Wan2.2-Fun-Reward-LoRAs
- Publisher
- Alibaba-PAI
- Task
- Text to video
- Modality
- Video
- Library
- videox_fun
- Parameters
- Not stated by the source
- Languages
- Not stated by the source
- Revision
- 98ccd8617f3101dc7ca388510dd7e65f5f42d91c
- First published
- 2025-09-03
- Last updated
- 2025-12-11
Files and Weights
7 files, 3.4 GB in total. The weights are 4 files totalling 3.4 GB in safetensors.
Weights4 files · 3.4 GB
Documentation2 files · 21.6 KB
Repository1 file · 1.5 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| Wan2.2-Fun-A14B-InP-high-noise-HPS2.1.safetensors | Weights | 858.5 MB | 35d014944141 |
| Wan2.2-Fun-A14B-InP-high-noise-MPS.safetensors | Weights | 858.5 MB | e9f969762891 |
| Wan2.2-Fun-A14B-InP-low-noise-HPS2.1.safetensors | Weights | 858.5 MB | e64ba8988172 |
| Wan2.2-Fun-A14B-InP-low-noise-MPS.safetensors | Weights | 858.5 MB | 0e6ac56c8906 |
| LICENSE.txt | Documentation | 11.3 KB | — |
| README.md | Documentation | 10.2 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 3.4 GB
Download from Alibaba-PAI
Released by Alibaba-PAI through its official repository on Hugging Face. Read the license.
Built From
- Derived from Wan-AI/Wan2.2-T2V-A14B
- Described by arXiv:2310.03739
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 3.4 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Wan2.2-Fun-Reward-LoRAs
Can I use Wan2.2-Fun-Reward-LoRAs commercially?
Yes. Wan2.2-Fun-Reward-LoRAs is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
Similar Models
A LoRA for MiniMax-H3 that renders joint video + synchronized stereo audio in as few as 4 sampling steps instead of the usual ~20 — a ~5× sampling speedup — and keeps getting better as you add steps. For most work, use minimaxh3turbov4step600ema.safetensors. It's the markedly better micro-detail (faces, fingers, fine texture), and the over-sharpening / plastic look of the earlier v1 (~850) line is fully resolved. v4 introduced a static-frame enhancement — a big win for static and small-motion content. The one trade-off shows up only at 4 steps with large, fast motion, where v4 can produce motion-smear / trailing ghosting (we're actively fixing this). Two things address it: - Use 6–8 steps.…
This repository contains MiniMax-H3 Turbo LoRAs converted and optimized for ComfyUI: These LoRAs accelerate MiniMax-H3 video and synchronized-audio generation by reducing the required number of sampling steps. Newly added LoRA, located in the experimental/ folder: Manual recommended sigmas: 3-step 1.0, 0.961165, 0.853333, 0.0 4-step 1.0, 0.970874, 0.907249, 0.640000, 0.0 Three LoRAs extracted from VDN-H3 8 step: The main 8-step LoRA works on both FL2VA and Ref2VA. If you are running a pruned base, choose the pruned version that corresponds to your base — the pruned versions need their own matching pruned base. Three dynamically resized BF16 LoRAs are now included. Their source weights were…
Sulphur 2 An uncensored video generation model based on LTX 2.3 supporting both t2v and i2v natively, as well as all of the other ltx 2.3 formats. Follow us on X Join our Discord Support the next version of the project, even just a few dollars would go a long way: Kofi To get started with the model, I recommend downloading either of the dev versions, (fp8mixed or bf16) and downloading the distill lora provided. By the way, I'm aware the workflows contain sulphurfinal right now, just use the lora or use the full models, don't use both at the same time. This model contains a prompt enhancer. The easiest way to get started with the prompt enhancer is by using it on lmstudio. The way to…
Wan2.1 is an open-source suite of video foundation models, compatible with consumer-grade GPUs, that excels in various video generation tasks like text-to-video, image-to-video, and video editing, even supporting visual text generation. Download models using huggingface-cli: You can also download directly from this page. This model is a derivative work of the original model licensed under the Apache 2.0 License, and is therefore distributed under the terms of the same license. Thanks to Patrick Gillespie for creating the ASCII text art tool used in this project https://patorjk.com/software/taag/ Wan-AI for the Wan model https://huggingface.co/Wan-AI/Wan2.1-VACE-1.3B…
Wan2.1 is an open-source suite of video foundation models, compatible with consumer-grade GPUs, that excels in various video generation tasks like text-to-video, image-to-video, and video editing, even supporting visual text generation. Download models using huggingface-cli: You can also download directly from this page. This model is a derivative work of the original model licensed under the Apache 2.0 License, and is therefore distributed under the terms of the same license. Thanks to Patrick Gillespie for creating the ASCII text art tool used in this project https://patorjk.com/software/taag/ Wan-AI for the Wan model https://huggingface.co/Wan-AI/Wan2.1-T2V-1.3B…