Aller au contenu
ModelPrésentéExterne
Télécharger

AnimeGen T2V A14B Lightning Int4

An MLX (Apple Silicon) conversion of aidealab/AnimeGen-T2V — an anime-specialized full fine-tune of Wan2.2-T2V-A14B (dual-expert MoE) — quantized to int4 and shipped as a pre-merged Lightning checkpoint for few-step,...

Sourcehuggingface.co/mlx-community/AnimeGen-T2V-A14B-Lightning-int4/tree/3cbba73faba2f480a5635171160e957a3603e57d
TypeOtherModèle de baseWan-AI/Wan2.2-T2V-A14BVersionRevision 3cbba73BClasse

Source du modèle

Type
Other
Taille du fichier
484 MB–10,6 GB
Modèle de base
Wan-AI/Wan2.2-T2V-A14B
Version
vRevision 3cbba73
Mis à jour
9 oct. 2026
Téléchargements
49
Source
Hugging Face

Description de la source

AnimeGen-T2V-A14B-Lightning-int4 (MLX)

An MLX (Apple Silicon) conversion of aidealab/AnimeGen-T2V — an anime-specialized full fine-tune of Wan2.2-T2V-A14B (dual-expert MoE) — quantized to int4 and shipped as a pre-merged Lightning checkpoint for few-step, CFG-free inference.

This variant is built for the Swift-MLX wan-core / Bernini-R runtime: it uses the exact flat, 1095-key A14B checkpoint layout that runtime loads, so the DiT experts drop in unchanged.

What's in here

Flat checkpoint (stock-Wan VAE + umT5 reused from Wan2.2-T2V-A14B, byte-identical):

FileWhat
high_noise_model.safetensorsfine-tuned high-noise expert, int4-quantized (attention + FFN linears)
low_noise_model.safetensorsfine-tuned low-noise expert, int4-quantized
vae.safetensors16-channel Wan VAE (stock Wan2.2)
t5_encoder.safetensorsumT5-XXL text encoder (stock Wan2.2)
config.jsonWan2.2-A14B config + quantization metadata

How it was made

  1. The two fine-tuned experts from aidealab/AnimeGen-T2V (original-Wan key naming, bf16) were converted 1:1 into the MLX A14B layout (sanitize_wan_transformer_weights; the key set is asserted equal to the base A14B contract, 1095 keys, 0 added / 0 dropped).
  2. AIdeaLab's acceleration recipe — the lightx2v Wan2.2-Lightning Wan2.2-T2V-A14B-4steps-lora-250928 LoRA — was merged offline per-expert with AIdeaLab's weighting (set_adapters([high, low], [2.0, 1.0]); kohya W += strength·(alpha/rank)·(up @ down)).
  3. The attention + FFN linears of both experts were quantized to int4 (group size 64); VAE + umT5 kept full precision.

Recommended inference recipe (from the AnimeGen model card)

  • Sampler: euler, CFG-free (guidance 1.0)
  • 8 steps, shift 3.0 (FlowMatchEuler)
  • 832×480 or 1280×720, frames 16·seconds + 1 (must be 4n+1), 16 fps

Footprint

int4 measured ~42 GB peak at 832×480 / 17 frames / 8 steps on an M-series GPU (fits a 64 GB Mac).

License & attribution

Apache-2.0. Derived from aidealab/AnimeGen-T2V (AIdeaLab, developed with GENIAC/METI support) and the lightx2v Lightning LoRA. This is an unofficial community MLX conversion; not affiliated with AIdeaLab.

Sources

1 source
Hugging FacePrincipalVersion 3cbba73faba2f480a5635171160e957a3603e57d
huggingface.co/mlx-community/AnimeGen-T2V-A14B-Lightning-int4/tree/3cbba73faba2f480a5635171160e957a3603e57d

Vérifié 9 oct.

Artefacts du modèle

4 artefacts

high_noise_model.safetensors

safetensors · 7,80 GB · SHA-256 20d3e34cb55e…1f2e · Hugging Face

Télécharger

low_noise_model.safetensors

safetensors · 7,80 GB · SHA-256 87342c3bb6e6…5f7b · Hugging Face

Télécharger

t5_encoder.safetensors

safetensors · 10,6 GB · SHA-256 e86ee4199903…3787 · Hugging Face

Télécharger

vae.safetensors

safetensors · 484 MB · SHA-256 977530e453db…7414 · Hugging Face

Télécharger

Extraits de sources

3 extraits
Readme31 juil.huggingface.co/mlx-community/AnimeGen-T2V-A14B-Lightning-int4

--- license: apache-2.0 base_model: - aidealab/AnimeGen-T2V - Wan-AI/Wan2.2-T2V-A14B tags: - text-to-video - anime - mlx - wan2.2 library_name: mlx --- # AnimeGen-T2V-A14B-Lightning-int4 (MLX) An **MLX (Apple Silicon)** conversion of [aidealab/AnimeGen-T2V](https://huggingface.co/aidealab/AnimeGen-T2V) — an anime-specialized **full fine-tune of Wan2.2-T2V-A14B** (dual-expert MoE) — quantized to **int4** and shipped as a pre-merged **Lightning** checkpoint for few-step, CFG-free inference. This variant is built for the Swift-MLX `wan-core` / Bernini-R runtime: it uses the exact flat, 1095-key A14B checkpoint layout that runtime loads, so the DiT experts drop in unchanged. ## What's in here Flat checkpoint (stock-Wan VAE + umT5 reused from Wan2.2-T2V-A14B, byte-identical): | File | What | |---|---| | `high_noise_model.safetensors` | fine-tuned high-noise expert, int4-quantized (attention + FFN linears) | | `low_noise_model.safetensors` | fine-tuned low-noise expert, int4-quantized | | `vae.safetensors` | 16-channel Wan VAE (stock Wan2.2) | | `t5_encoder.safetensors` | umT5-XXL text encoder (stock Wan2.2) | | `config.json` | Wan2.2-A14B config + `quantization` metadata | ## How it was made 1. The two fine-tuned experts from `aidealab/AnimeGen-T2V` (original-Wan key naming, bf16) were converted 1:1 into the MLX A14B layout (`sanitize_wan_transformer_weights`; the key set is asserted equal to the base A14B contract, 1095 keys, 0 added / 0 dropped). 2. AIdeaLab's acceleration recipe — the [lightx2v Wan2.2-Lightning](https://huggingface.co/lightx2v/Wan2.2-Lightning) `Wan2.2-T2V-A14B-4steps-lora-250928` LoRA — was **merged offline** per-expert with AIdeaLab's weighting (`set_adapters([high, low], [2.0, 1.0])`; kohya `W += strength·(alpha/rank)·(up @ down)`). 3. The attention + FFN linears of both experts were quantized to int4 (group size 64); VAE + umT5 kept full precision. ## Recommended inference recipe (from the AnimeGen model card) - Sampler: **euler**, **CFG-free** (guidance 1.0) - **8 steps**, **shift 3.0** (FlowMatchEuler) - 832×480 or 1280×720, frames `16·seconds + 1` (must be 4n+1), 16 fps ## Footprint int4 measured **~42 GB peak** at 832×480 / 17 frames / 8 steps on an M-series GPU (fits a 64 GB Mac). ## License & attribution Apache-2.0. Derived from `aidealab/AnimeGen-T2V` (AIdeaLab, developed with GENIAC/METI support) and the...

Note9 oct.huggingface.co/mlx-community/AnimeGen-T2V-A14B-Lightning-int4

Source context: 49 downloads · 0 likes · Pipeline text-to-video · Library mlx · Repo mlx-community/AnimeGen-T2V-A14B-Lightning-int4

Note31 juil.huggingface.co/mlx-community/AnimeGen-T2V-A14B-Lightning-int4

Source context: 65 downloads · 0 likes · Pipeline text-to-video · Library mlx · Repo mlx-community/AnimeGen-T2V-A14B-Lightning-int4