Commits · c335fdf2000d2e45640c01c4e89ef88c131dda53 · chenpangpang / ComfyUI · GitLab

01 Sep, 2023 2 commits
- Remove xformers related print. · 0e3b6411
  comfyanonymous authored Sep 01, 2023
  
  0e3b6411
- Fix controlnet bug. · 5c363a9d
  comfyanonymous authored Sep 01, 2023
  
  5c363a9d
31 Aug, 2023 2 commits
- Fix controlnet issue. · cfe1c54d
  comfyanonymous authored Aug 31, 2023
  
  cfe1c54d
- It doesn't make sense for c_crossattn and c_concat to be lists. · 1c012d69
  comfyanonymous authored Aug 31, 2023
  
  1c012d69
30 Aug, 2023 3 commits
- Clean up DiffusersLoader node. · 7e941f9f
  comfyanonymous authored Aug 30, 2023
  
  7e941f9f
- Fix error message in model_patcher.py · 18617967
  Simon Lui authored Aug 30, 2023
```
Found while tinkering.
```
  18617967
- Fix "Load Checkpoint with config" node. · fe4c0740
  comfyanonymous authored Aug 29, 2023
  
  fe4c0740
29 Aug, 2023 3 commits
- Support SDXL t2i adapters with 3 channel input. · f2f5e5dc
  comfyanonymous authored Aug 29, 2023
  
  f2f5e5dc
- Move beta_schedule to model_config and allow disabling unet creation. · 15adc369
  comfyanonymous authored Aug 29, 2023
  
  15adc369
- Remove optimization that caused border. · bed116a1
  comfyanonymous authored Aug 29, 2023
  
  bed116a1
28 Aug, 2023 9 commits
- No need to check filename extensions to detect shuffle controlnet. · 65cae62c
  comfyanonymous authored Aug 28, 2023
  
  65cae62c
- Put clip vision outputs on the CPU. · 4e89b2c2
  comfyanonymous authored Aug 28, 2023
  
  4e89b2c2
- Load clipvision model to GPU for faster performance. · a094b45c
  comfyanonymous authored Aug 28, 2023
  
  a094b45c
- Text encoder should initially load on the offload_device not the regular. · 1300a1bb
  comfyanonymous authored Aug 28, 2023
  
  1300a1bb
- Move ModelPatcher to model_patcher.py · f92074b8
  comfyanonymous authored Aug 28, 2023
  
  f92074b8
- Implement loras with norm keys. · 4798cf5a
  comfyanonymous authored Aug 28, 2023
  
  4798cf5a
- Enable bf16-vae by default on ampere and up. · b8c7c770
  comfyanonymous authored Aug 27, 2023
  
  b8c7c770
- Fallback to slice attention if xformers doesn't support the operation. · 1c794a21
  comfyanonymous authored Aug 27, 2023
  
  1c794a21
- Make --bf16-vae work on torch 2.0 · d935ba50
  comfyanonymous authored Aug 27, 2023
  
  d935ba50
26 Aug, 2023 1 commit
- Fix lowvram model merging. · a57b0c79
  comfyanonymous authored Aug 26, 2023
  
  a57b0c79
25 Aug, 2023 4 commits
- The new smart memory management makes this unnecessary. · f72780a7
  comfyanonymous authored Aug 25, 2023
  
  f72780a7
- Move controlnet code to comfy/controlnet.py · c77f02e1
  comfyanonymous authored Aug 25, 2023
  
  c77f02e1
- Move lora code to comfy/lora.py · 15a7716f
  comfyanonymous authored Aug 25, 2023
  
  15a7716f
- Move text_projection to base clip model. · ec96f6d0
  comfyanonymous authored Aug 24, 2023
  
  ec96f6d0
24 Aug, 2023 6 commits
- Code cleanups. · 30eb92c3
  comfyanonymous authored Aug 24, 2023
  
  30eb92c3
- Try to free enough vram for control lora inference. · 51dde87e
  comfyanonymous authored Aug 24, 2023
  
  51dde87e
- Fix potential issue with text projection matrix multiplication. · e3d0a9a4
  comfyanonymous authored Aug 24, 2023
  
  e3d0a9a4
- Always shift text encoder to GPU when the device supports fp16. · cc44ade7
  comfyanonymous authored Aug 23, 2023
  
  cc44ade7
- Even with forced fp16 the cpu device should never use it. · a6ef08a4
  comfyanonymous authored Aug 23, 2023
  
  a6ef08a4
- Initialize text encoder to target dtype. · 00c0b2c5
  comfyanonymous authored Aug 23, 2023
  
  00c0b2c5
23 Aug, 2023 1 commit
- Save memory by storing text encoder weights in fp16 in most situations. · f081017c
  comfyanonymous authored Aug 23, 2023
```
Do inference in fp32 to make sure quality stays the exact same.
```
  f081017c
22 Aug, 2023 4 commits
- All resolutions now work with t2i adapter for SDXL. · afcb9cb1
  comfyanonymous authored Aug 22, 2023
  
  afcb9cb1
- T2I adapter SDXL. · 85fde89d
  comfyanonymous authored Aug 22, 2023
  
  85fde89d
- Controlnet/t2iadapter cleanup. · cf5ae469
  comfyanonymous authored Aug 21, 2023
  
  cf5ae469
- Fix control lora not working in fp32. · 763b0cf0
  comfyanonymous authored Aug 21, 2023
  
  763b0cf0
21 Aug, 2023 2 commits
- Fix ControlLora on lowvram. · 199d7336
  comfyanonymous authored Aug 21, 2023
  
  199d7336
- Remove autocast from controlnet code. · d08e53de
  comfyanonymous authored Aug 20, 2023
  
  d08e53de
20 Aug, 2023 3 commits
- Small cleanups. · 0d7b0a4d
  comfyanonymous authored Aug 20, 2023
  
  0d7b0a4d
- Further tuning and fix mem_free_total. · 92254659
  Simon Lui authored Aug 19, 2023
  
  92254659
- Add ipex optimize and other enhancements for Intel GPUs based on recent memory changes. · 2c096e42
  Simon Lui authored Aug 17, 2023
  
  2c096e42