# Inference - [X] Add DGX Spark support hao-ai-lab/FastVideo#1447 - [ ] Streamline and improve documentation for DGX Spark - [ ] Improve nvfp4/fp8 inference API and example scripts @kevin314 [#1496](<https://github.com/hao-ai-lab/FastVideo/issues/1496>) - [ ] batching support hao-ai-lab/FastVideo#1453 @macthecadillac - [ ] device specific optimal example scripts (eg 5090, 3090, GB200, etc) - [ ] UI frontend for inference ## Model support Tracking issue: [https://github.com/hao-ai-lab/FastVideo/issues/1407](<https://github.com/hao-ai-lab/FastVideo/issues/1407>) ### Image Models - [X] Flux 1 [#1321](<https://github.com/hao-ai-lab/FastVideo/issues/1321>) - [X] Flux 2 [#1349](<https://github.com/hao-ai-lab/FastVideo/issues/1349>) - [X] GLM-Image hao-ai-lab/FastVideo#1030 ### Video Models - [X] Kandinsky-5 T2V/I2V @leffff [#1471](<https://github.com/hao-ai-lab/FastVideo/issues/1471>) - [ ] Cosmos3 @SolitaryThinker - [ ] Add Lingbot-Video @Davids048 [#1600](<https://github.com/hao-ai-lab/FastVideo/issues/1600>) ### AR models - [X] DreamX-world 5B @Suckl [#1538](<https://github.com/hao-ai-lab/FastVideo/issues/1538>) - [X] Add Lingbot-World @Davids048 hao-ai-lab/FastVideo#1579 - [ ] Stable-Video-Infinity @H1yori233 hao-ai-lab/FastVideo#1344 ## Dreamverse - [ ] Dreamverse i2v Support @SolitaryThinker - [ ] Support Cosmos 2.5 Predict 2B @jolettypaulraj - [ ] Add 5090 support # Training - [ ] Add Muon optimizer @alexzms hao-ai-lab/FastVideo#1519 - [ ] Add Kandinsky FT @leffff - [ ] Add LTX2.3 FT @SolitaryThinker - [ ] webui frontend for FT ## Distillation - [X] anyflow hao-ai-lab/FastVideo#1371 - [X] Teacher forcing and causal consistency distillation @H1yori233 [#1505](<https://github.com/hao-ai-lab/FastVideo/issues/1505>) - [X] ReROPE @H1yori233 hao-ai-lab/FastVideo#1454 - [ ] Add TDM distillation [https://arxiv.org/pdf/2503.06674](<https://arxiv.org/pdf/2503.06674>) - [ ] i2v distillation @alexzms @Satyam-53 - [ ] kandinsky QAD support @leffff hao-ai-lab/FastVideo#1601 - [ ] LTX2.3 QAD support @SolitaryThinker - [ ] Optimize QAD MFU (in progress) ## RL - [ ] Add DiffusionNFT Wan Video RL training @Abecid [#1476](<https://github.com/hao-ai-lab/FastVideo/issues/1476>) - [ ] PromptRL @Davids048 # Kernels - [ ] Flashinfer Attention backend @SolitaryThinker - [ ] Optimize nvpf4 QAT kernel for DGX Spark - [ ] Kernel fusion support for popular models - [ ] Wan - [ ] LTX - [ ] Cosmos - [ ] Kandinsky ## CI / Perf Optimization Tracking Issue: hao-ai-lab/FastVideo#1374 - [ ] Land Performance CI phase 1 @Satyam-53 - [ ] Baseline and coverage for Wan, LTX, and Cosmos model @Satyam-53 - [X] Spin up the perf dashboard on a separate CPU instance @Satyam-53
Inference
Model support
Tracking issue: #1407
Image Models
Video Models
AR models
Dreamverse
Training
Distillation
RL
Kernels
CI / Perf Optimization
Tracking Issue: #1374