Repository navigation

All pull requests

Labels (9)Milestones (0)New pull request

Open

(24)

Closed

(58)

fix(pipelines): seed prompt enhancement with the request seed and make it reproducible
#324Christopher (christopher5106) opened Sep 29, 2026
fix(pipelines): inference_mode for library calls; CPU-first safetensors load
#323Ray Winkelman (rwinkelman) opened Sep 28, 20263 tasks
ltx-trainer: store caption latents trimmed to their real tokens
#321Shaul Shvimmer (CallShaul) opened Sep 24, 2026
loader: fail when a completed build retains meta tensors
#317liminfei-amd (liminfei-amd) opened Sep 22, 2026
fix(kernels): build on rolling-release toolchains (CUDA_PATH veto, POSIX macro clash)
#313ManuelFCastillo opened Sep 14, 2026
docs: OffloadMode.NONE VRAM figure applies to the 19B LTX-2, not LTX-2.5
#312Keith (stayinalive181) opened Sep 10, 2026
Fix dependency resolution on macOS / Apple Silicon
#298medyadanhasan opened Aug 24, 2026
Handle transposed fp8 weights in LoRA fusion
#293masahiroteraoka opened Aug 20, 2026
Reference conditioning: RoPE layouts and per-source rotary phase
#289NRDX (alisson-anjos) opened Aug 17, 2026
pipelines: forward --quantization to the multi-GPU runners
#280Boning (wwelder) opened Aug 13, 2026
Preserve fractional frame rates when encoding SDR video
#279Syed Osama Ali Shah (Osamaali313) opened Aug 12, 2026
1
fix(trainer): restore valid ruff target-version
#265blue (Blueyyds) opened Aug 2, 2026
fix(kernels): don't fail all2all build on CUDA 12.8 deprecation warnings
#264blue (Blueyyds) opened Aug 2, 2026
fix(trainer): correct LR schedule, final checkpoint and STG under DDP
#263blue (Blueyyds) opened Aug 2, 2026
Ampere (sm_86) compatibility for fp8-cast LoRA fusion and mono-audio input in a2vid
#253Larry Bowden (fastflair) opened Jul 11, 2026
README: Update outdated url
#238Giuseppe Silvano (giusilvano) opened Jun 18, 2026
Limit seq_length to max_tokens in timestep_samplers
#230lsl0007 opened Jun 8, 2026
perf(trainer): add tile_batch_size, cut VAE calls up to 15x in tiled encoding
#215Vittoria Lanzo (VittoriaLanzo) opened May 14, 20265 tasks done
1
Add keyframe-to-video training strategy
#209yizhiqianbi opened May 4, 2026
Free Gemma text encoder after caption embedding to reduce VRAM during training
#85bruefire (bruefire) opened Jan 21, 2026
Bugfix: fix memory allocation issues during multi-GPU training
#67LittleNyima (LittleNyima) opened Jan 14, 2026
8Gb Vram optimized, Win11 support + Web UI + optimized transformer +40% more frames on 8Gb
#66nalexand opened Jan 14, 2026
7
feat: Add deterministic inference pipeline for reproducible video generation
#28q5sys (JT) (q5sys) opened Jan 8, 2026
1
Small README edits
#20Jack Saunders (jsaunders909) opened Jan 7, 2026

Pull requests