mirror of https://github.com/vladmandic/automatic
3.8 KiB
3.8 KiB
TODO
Main ToDo list can be found at GitHub projects
Future Candidates
- Remote TE
- Canvas
- Refactor: Modular pipelines and guiders
- Refactor: Sampler options
- Refactor: GGUF
- Feature: LoRA add OMI format support for SD35/FLUX.1
- Video Core: API
- Video LTX: TeaCache and others, API, Conditioning preprocess Video: LTX API
Under Consideration
- Inf-DiT
- X-Omni
- DiffSynth Studio
- IPAdapter negative guidance
- IPAdapter composition
- STG
- SmoothCache
- MagCache
- Nunchaku PulID
- Dream0 guidance
- SUPIR upscaler
- ByteDance OneReward
- ByteDance USO
- Remove:
Agent Scheduler - Remove:
CodeFormer - Remove:
GFPGAN - ModernUI: Lite vs Expert mode
- Engine: TensorRT acceleration
New models
- HunyuanImage
- Phantom HuMo
- Lumina-DiMOO
- Wan2.2-Animate-14B
- Magi(https://github.com/huggingface/diffusers/pull/11713)
- SEVA
- Ming
- Liquid
- Step1X
- LucyEdit
- SD3 UltraEdit
- WAN2GP
- SelfForcing
- DiffusionForcing
- LanDiff
- HunyuanCustom
- HunyuanAvatar
- WAN-CausVid
- WAN-CausVid-Plus t2v
- WAN-StepDistill
Code TODO
pnpm lint | grep W0511 | awk -F'TODO ' '{print "- "$NF}' | sed 's/ (fixme)//g' | sort
- control: support scripts via api
- fc: autodetect distilled based on model
- fc: autodetect tensor format based on model
- hypertile: vae breaks when using non-standard sizes
- install: enable ROCm for windows when available
- loader: load receipe
- loader: save receipe
- lora: add other quantization types
- lora: add t5 key support for sd35/f1
- lora: maybe force imediate quantization
- model load: force-reloading entire model as loading transformers only leads to massive memory usage
- model load: implement model in-memory caching
- modernui: monkey-patch for missing tabs.select event
- modules/lora/lora_extract.py:188:9: W0511: TODO: lora: support pre-quantized flux
- processing: remove duplicate mask params
- resize image: enable full VAE mode for resize-latent