Skip to page content
SAVRN
Search Plan a project

Open-weight model · Image to video

flux3-agibot-looped

by Song Jinyan0924/flux3-agibot-looped

flux3-agibot-looped is an open-weight model for image to video from Song, released under other. Its published files total 555.5 GB. It draws 58 downloads a month.

Looped Diffusion Transformer (Chng et al., 2026, arXiv:2609.40305) retrofit of Jinyan0924/flux3-agibot-videoft (itself a full fine-tune of black-forest-labs/flux-3-action-base): the 33 joint blocks are split 12 / 9 / 12, the middle 9 are run N = 4 times with…

Parameters—
Context—
Weights222.1 GB
Licenseother
AccessOpen weights
Monthly Downloads58

Model Card

Looped Diffusion Transformer (Chng et al., 2026, arXiv:2609.40305) retrofit of Jinyan0924/flux3-agibot-videoft (itself a full fine-tune of black-forest-labs/flux-3-action-base): the 33 joint blocks are split 12 / 9 / 12, the middle 9 are run N = 4 times with head-wise gated attention, and every exit is trained with deep supervision. Task: k past head-camera frames + caption → 33 future frames at 15 fps, Val loss (EMA, N = 4, lower is better; k = number of context frames): r1 step 2500 k1 0.0787 / k9 0.0654; r2 step 4000 k1 0.0783 / k9 0.0651 (the step-4500 val pass was interrupted by the crash that ended the run, so the last val point is step 4000). Test-split PSNR / SSIM / LPIPS / FVD and…

Excerpt from the card by Song, licensed other.

Configuration

Stored precision
float32

Identity and Version

Repository
Jinyan0924/flux3-agibot-looped
Publisher
Song
Task
Image to video
Modality
Other
Library
Not stated by the source
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
93a12b2b2325b246e663a2742033f2fb63247330
First published
2026-10-14
Last updated
2026-10-17

Files and Weights

87 files, 555.5 GB in total. The weights are 30 files totalling 222.1 GB in pt, safetensors.

Weights30 files · 222.1 GB
Configuration13 files · 38.3 KB
Documentation3 files · 23.2 KB
Other30 files · 333.5 GB
Repository11 files · 4.1 MB
Every file
FileTypeSizeSHA-256
checkpoints/step-4500/ema_cpu_0p999/rank-0000.ptWeights11.1 GB 3b1b56225122
checkpoints/step-4500/ema_cpu_0p999/rank-0001.ptWeights11.1 GB 4cd5f5824fe4
checkpoints/step-4500/rng/rank-0000.ptWeights13.9 KB ee1bc027d8c1
checkpoints/step-4500/rng/rank-0001.ptWeights13.8 KB b2ee17f76994
ema_loop2500_step-2500.safetensorsWeights11.1 GB a56d73ed942e
ema_loop4500_step-4500.safetensorsWeights11.1 GB 5bef4fea2531
ema_step-4000.safetensorsWeights11.1 GB 4f8e90bf06a9
ema_step-4500.safetensorsWeights11.1 GB 5bef4fea2531
r1/checkpoints/step-2500/ema_cpu_0p999/rank-0000.ptWeights11.1 GB fe095b5f6d3b
r1/checkpoints/step-2500/ema_cpu_0p999/rank-0001.ptWeights11.1 GB 02b60ef68b69
r1/checkpoints/step-2500/rng/rank-0000.ptWeights13.9 KB 1f47b74d21b6
r1/checkpoints/step-2500/rng/rank-0001.ptWeights13.8 KB 431907c20f95
random_depth_n1-8/checkpoints/step-2000/ema_cpu_0p999/rank-0000.ptWeights11.1 GB bf99bc11388e
random_depth_n1-8/checkpoints/step-2000/ema_cpu_0p999/rank-0001.ptWeights11.1 GB 8f9db8c80fc5
random_depth_n1-8/checkpoints/step-2000/rng/rank-0000.ptWeights13.9 KB 5663b3525a43
random_depth_n1-8/checkpoints/step-2000/rng/rank-0001.ptWeights13.8 KB 42bfd4876d37
random_depth_n1-8/checkpoints/step-2500/ema_cpu_0p999/rank-0000.ptWeights11.1 GB 147d3781e395
random_depth_n1-8/checkpoints/step-2500/ema_cpu_0p999/rank-0001.ptWeights11.1 GB 614a88223d59
random_depth_n1-8/checkpoints/step-2500/rng/rank-0000.ptWeights13.9 KB 1f47b74d21b6
random_depth_n1-8/checkpoints/step-2500/rng/rank-0001.ptWeights13.8 KB 431907c20f95
random_depth_n1-8/ema_step-1000.safetensorsWeights11.1 GB 8dc707a6f180
random_depth_n1-8/ema_step-1500.safetensorsWeights11.1 GB 0a6d0cd6bc0c
random_depth_n1-8/ema_step-2000.safetensorsWeights11.1 GB dd13edfb8238
random_depth_n1-8/ema_step-2500.safetensorsWeights11.1 GB 641c34bab731
random_depth_n1-8/ema_step-500.safetensorsWeights11.1 GB 6f7925a95b8a
random_depth_n1-8_long/checkpoints/step-2500/ema_cpu_0p999/rank-0000.ptWeights11.1 GB a91cb9343983
random_depth_n1-8_long/checkpoints/step-2500/ema_cpu_0p999/rank-0001.ptWeights11.1 GB 4cf323c36a8f
random_depth_n1-8_long/checkpoints/step-2500/rng/rank-0000.ptWeights13.9 KB ac703afe3258
random_depth_n1-8_long/checkpoints/step-2500/rng/rank-0001.ptWeights13.8 KB 306dd377f5ec
random_depth_n1-8_long/ema_step-2500.safetensorsWeights11.1 GB 4fc6fca5da6c
checkpoints/step-4500/config.jsonConfiguration2.0 KB —
checkpoints/step-4500/state.jsonConfiguration4.4 KB —
config.jsonConfiguration2.0 KB —
r1/checkpoints/step-2500/config.jsonConfiguration2.0 KB —
r1/checkpoints/step-2500/state.jsonConfiguration3.9 KB —
random_depth_n1-8/checkpoints/step-2000/config.jsonConfiguration2.2 KB —
random_depth_n1-8/checkpoints/step-2000/state.jsonConfiguration4.1 KB —
random_depth_n1-8/checkpoints/step-2500/config.jsonConfiguration2.2 KB —
random_depth_n1-8/checkpoints/step-2500/state.jsonConfiguration4.1 KB —
random_depth_n1-8/config.jsonConfiguration2.2 KB —
random_depth_n1-8_long/checkpoints/step-2500/config.jsonConfiguration2.2 KB —
random_depth_n1-8_long/checkpoints/step-2500/state.jsonConfiguration4.7 KB —
random_depth_n1-8_long/config.jsonConfiguration2.2 KB —
LICENSE.mdDocumentation19.6 KB —
NOTICEDocumentation993 B —
README.mdDocumentation2.6 KB —
checkpoints/step-4500/COMPLETEOther5 B —
checkpoints/step-4500/model/__0_0.distcpOther11.1 GB eb5002adc546
checkpoints/step-4500/model/__1_0.distcpOther11.1 GB b9357dd2271a
checkpoints/step-4500/optimizer/__0_0.distcpOther22.2 GB 3d0a69507c06
checkpoints/step-4500/optimizer/__1_0.distcpOther22.2 GB cd8e62be5060
eval/val.jsonlOther2.8 KB —
metrics.jsonlOther200.5 KB —
r1/checkpoints/step-2500/COMPLETEOther5 B —
r1/checkpoints/step-2500/model/__0_0.distcpOther11.1 GB 29d08db2cabf
r1/checkpoints/step-2500/model/__1_0.distcpOther11.1 GB 08cb633e97fe
r1/checkpoints/step-2500/optimizer/__0_0.distcpOther22.2 GB 0929bcfc5b24
r1/checkpoints/step-2500/optimizer/__1_0.distcpOther22.2 GB 133b54240ad3
random_depth_n1-8/checkpoints/step-2000/COMPLETEOther5 B —
random_depth_n1-8/checkpoints/step-2000/model/__0_0.distcpOther11.1 GB dcfd7a9c77d8
random_depth_n1-8/checkpoints/step-2000/model/__1_0.distcpOther11.1 GB 350a6796e9bb
random_depth_n1-8/checkpoints/step-2000/optimizer/__0_0.distcpOther22.2 GB 9643acc4159e
random_depth_n1-8/checkpoints/step-2000/optimizer/__1_0.distcpOther22.2 GB 3742b7813e9c
random_depth_n1-8/checkpoints/step-2500/COMPLETEOther5 B —
random_depth_n1-8/checkpoints/step-2500/model/__0_0.distcpOther11.1 GB f53061da3628
random_depth_n1-8/checkpoints/step-2500/model/__1_0.distcpOther11.1 GB 8f26a83a752c
random_depth_n1-8/checkpoints/step-2500/optimizer/__0_0.distcpOther22.2 GB 2c1b850d4de4
random_depth_n1-8/checkpoints/step-2500/optimizer/__1_0.distcpOther22.2 GB acb82915bece
random_depth_n1-8/eval/val.jsonlOther1.6 KB —
random_depth_n1-8/metrics.jsonlOther101.9 KB —
random_depth_n1-8_long/checkpoints/step-2500/COMPLETEOther5 B —
random_depth_n1-8_long/checkpoints/step-2500/model/__0_0.distcpOther11.1 GB 1adb441fd11e
random_depth_n1-8_long/checkpoints/step-2500/model/__1_0.distcpOther11.1 GB 66eff39dbf4c
random_depth_n1-8_long/checkpoints/step-2500/optimizer/__0_0.distcpOther22.2 GB b53ba4dc8184
random_depth_n1-8_long/checkpoints/step-2500/optimizer/__1_0.distcpOther22.2 GB b83bb5a3593f
random_depth_n1-8_long/metrics.jsonlOther20.3 KB —
.gitattributesRepository6.1 KB —
checkpoints/step-4500/model/.metadataRepository214.6 KB 09387406a578
checkpoints/step-4500/optimizer/.metadataRepository604.6 KB af12bb796a13
r1/checkpoints/step-2500/model/.metadataRepository214.6 KB cfba887c5778
r1/checkpoints/step-2500/optimizer/.metadataRepository604.6 KB 21b8f4b836a7
random_depth_n1-8/checkpoints/step-2000/model/.metadataRepository214.6 KB fa1f21824906
random_depth_n1-8/checkpoints/step-2000/optimizer/.metadataRepository604.6 KB 9a2dd65d41cb
random_depth_n1-8/checkpoints/step-2500/model/.metadataRepository214.6 KB 75802a76eaa8
random_depth_n1-8/checkpoints/step-2500/optimizer/.metadataRepository604.6 KB 0e8af4cbf512
random_depth_n1-8_long/checkpoints/step-2500/model/.metadataRepository214.6 KB acfc88edf12b
random_depth_n1-8_long/checkpoints/step-2500/optimizer/.metadataRepository604.7 KB 034ea6e36b8f

License and Download

License
other
Access
Open weights, no gate
Download size
222.1 GB
Download from Song

Released by Song through its official repository on Hugging Face.

Built From

  • Derived from black-forest-labs/flux-3-action-base
  • Described by arXiv:2609.40305

Memory Requirements

PrecisionWeights in memory
As published222.1 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About flux3-agibot-looped

What license is flux3-agibot-looped released under?

other, as its publisher declares it. Read the license text before commercial use.

Similar Models

Model · Image to video

Wan_2.2_ComfyUI_Repackaged

Comfy Org

Repackaged model files for ComfyUI. - https://huggingface.co/nvidia/ChronoEdit-14B-Diffusers - https://huggingface.co/Wan-AI/Wan2.2-Animate-14B - https://huggingface.co/alibaba-pai/Wan2.2-Fun-A14B-Control-Camera - https://huggingface.co/alibaba-pai/Wan2.2-Fun-A14B-Control - https://huggingface.co/alibaba-pai/Wan2.2-Fun-5B-Control - https://huggingface.co/alibaba-pai/Wan2.2-Fun-5B-InP - https://huggingface.co/alibaba-pai/Wan2.2-Fun-A14B-InP - https://huggingface.co/alibaba-pai/Wan2.2-VACE-Fun-A14B - https://huggingface.co/Wan-AI/Wan2.2-I2V-A14B - https://huggingface.co/Wan-AI/Wan2.2-S2V-14B - https://huggingface.co/Wan-AI/Wan2.2-T2V-A14B - https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B…

Open weights apache-2.0 diffusion-single-file
View model

Model · Image to video

LTX-2.5

LTX.io

LTX-2.5 is an open world model with open weights, built for local execution and fine-tuning. Its established use is generating synchronized, high-fidelity video and audio from text, image, and video inputs; applicability to emerging domains such as robotics and physical AI is developing. Full control and customization — self-host on your own infrastructure. No per-generation billing, no per-seat lock-in, no forced API dependency. Revenue is measured across the whole entity, including subsidiaries and affiliates under common control. The full, binding terms live in LICENSE. - Native multishot generation — generate connected scenes in a single pass: multiple shots that hold character…

Access requested at publisher other diffusion-single-file
View model

Model · Image to video

LTX-2.5

Comfyicu

LTX-2.5 is an open world model with open weights, built for local execution and fine-tuning. Its established use is generating synchronized, high-fidelity video and audio from text, image, and video inputs; applicability to emerging domains such as robotics and physical AI is developing. Full control and customization — self-host on your own infrastructure. No per-generation billing, no per-seat lock-in, no forced API dependency. Revenue is measured across the whole entity, including subsidiaries and affiliates under common control. The full, binding terms live in LICENSE. - Native multishot generation — generate connected scenes in a single pass: multiple shots that hold character…

Open weights other diffusion-single-file
View model

Model · Image to video

Minimax-h3-Turbo

Lightx2v

Please check our repository or the LightX2V MiniMax-H3 examples to reproduce the results. Please check the model specifications for more details. Try the MiniMax-H3 Turbo LoRA directly in LightX2V Studio: The Studio currently uses the FL2V 8-step v1.0 768p LoRA, which provides improved video and audio generation quality with 8-step inference. Integrate MiniMax-H3 Turbo into your application through the LightX2V API

Open weights apache-2.0 diffusers
View model

Model · Image to video

MiniMax-H3-GGUF

Jay

This repository (Abiray/MiniMax-H3-GGUF) provides GGUF quantized versions and necessary component files for the MiniMax H3 model. MiniMax H3 is a general-purpose, omni-modal generative system that supports unified understanding of multimodal contexts composed of text, images, video, and audio. It can generate video with native stereo audio at resolutions up to 2K and durations of up to 15 seconds. If you are looking for a smaller model with the same great quality that fits better on consumer-tier GPUs, please check out the MiniMax-H3-Pruned-GGUF repository. The pruned architecture is compressed down to 8.9 GB – 21.6 GB, bringing MiniMax H3 execution directly to consumer hardware. This…

Open weights other
View model

Model · Image to video

LTX-2.3

LTX.io

This model card focuses on the LTX-2.3 model, which is a significant update to the LTX-2 model with improved audio and visual quality as well as enhanced prompt adherence. LTX-2 was presented in the paper LTX-2: Efficient Joint Audio-Visual Foundation Model. If you want to dive in right to the code - it is available here. LTX-2.3 is a DiT-based audio-video foundation model designed to generate synchronized video and audio within a single model. It brings together the core building blocks of modern video generation, with open weights and a focus on practical, local execution. LTX-2.3 is accessible right away via the API Playground. You can use the models - full, distilled, upscalers and any…

Open weights other diffusers
View model