nodes/city96-ComfyUI_ExtraModels
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
115 Commits 2 Branches 0 Tags
d96f63d5ed1798462d07a0a0ec2ed63c021c5f0b
Commit Graph
38 Commits
Author SHA1 Message Date
City c6527a45f1 Set t5_attention_mask for new comfy versions 
#79
 2024-10-24 21:00:57 +02:00
City 193c2fd9d3 Remove unused arg 
#80
 2024-08-07 13:42:31 +02:00
City 6fb8649168 PixArt: fix OneTrainer LoRA 2024-07-27 17:52:02 +02:00
City 1db046de78 PixArt: Update import for SD3 clip 
#76
 2024-07-17 00:22:24 +02:00
a-One-Fan 8cff976421 Intel support for Pixart-Sigma (#64) 
* Fix IPEX attn_drop

* IPEX SDPA 4GB fix

Alchemist GPUs lack the 64 bit emulation needed for >4GB in this case. Attention slicing workaround copied from https://github.com/vladmandic/automatic/blob/master/modules/intel/ipex/attention.py

* Do 4GB fix only when necessary

~20% performance boost to not use the 4GB fix when not necessary

* Move attention.py inside utils

* Better 4GB estimation

* Remove unnecessary warning
 2024-07-06 20:26:29 +02:00
frutiemax92 e866b5b6cb Convert some missing layers (#72) 2024-07-02 01:23:52 +02:00
gchapman 46e4fc8ce3 Add new 900M config for the traditional loader, might be better to have depth selectable? 2024-06-23 19:48:34 +01:00
gchapman 8f42946305 Remove override as the new simple loader works it out for us. 2024-06-23 19:44:32 +01:00
codeMODE 7d0cb8fe31 Merge branch 'city96:main' into get-depth-programatically 2024-06-23 19:43:25 +01:00
gchapman 712f57c915 Use different keys for detection between original model and converted model 2024-06-19 09:45:33 +01:00
City de52d3aa45 PixArt auto checkpoint loader 
Needs better logic for diffusers weights
 2024-06-19 03:03:43 +02:00
City 2ad0b09cb9 PixArt allow on-the-fly PE interpolation 2024-06-19 02:26:01 +02:00
gchapman 91340dedc7 Fix loader depth 2024-06-18 23:56:10 +01:00
gchapman 5b082f424d Revert "Use correct key to get the number of layers, I accidentally used a layer with a 'final' layer giving me an out by one error" 
This reverts commit e43ecbeb64.
 2024-06-18 23:42:08 +01:00
gchapman e43ecbeb64 Use correct key to get the number of layers, I accidentally used a layer with a 'final' layer giving me an out by one error 2024-06-18 23:24:30 +01:00
gchapman 1d6c291e75 Minor optimization to stop a double .to() being needed 2024-06-18 21:45:10 +01:00
gchapman 03a81eddd5 Get the depth of the model by counting layers to allow for models with more depth. This allows for deeper models to be created. Example: https://huggingface.co/ptx0/pixart-reality-mix which is a 900M model 2024-06-18 21:44:42 +01:00
City 9a4629aba7 PixArt SD3 node prompt weight fix 
#61
 2024-06-16 02:51:21 +02:00
City c174662c5b Allow PixArt to use SD3 clip 2024-06-14 16:32:05 +02:00
City 9da9978f6c PixArt LoRA changes 
#60
 2024-06-12 01:04:22 +02:00
City 6867aacb26 PixArt fix float64 conversion error 
#48
 2024-05-25 18:09:56 +02:00
gchapman 0f888945f5 * Fix SDP attention mask by creating a diagonal mask in the right shape. 2024-05-16 09:23:16 +01:00
GavChap 3824b0d7b9 Update PixArt_blocks.py 
Add dropout_p to SDP attention
 2024-05-11 17:00:48 +01:00
City d52d732677 PixArt 2K config 
#23
 2024-04-23 18:31:40 +02:00
City 2f2797e502 PixArt Allow missing state dict keys 
Fixes #21
 2024-04-21 21:12:15 +02:00
City 5e3886c094 PixArt Sigma 1024 2024-04-09 23:48:34 +02:00
City c325dec0d3 PixArt sigma 
Mentioned in #19
Also removes old example sampler since updating it would be too much overhead.
Pytorch attention is still wonky.
 2024-04-07 17:55:33 +02:00
City 18fb3f5674 Quick fix for new comfy version 2024-04-06 12:05:04 +02:00
City 8befcdd9d6 PixArt controlnet test 2024-02-28 19:05:08 +01:00
City d2488d73f6 Quick FP8 fix. 2024-01-28 17:47:02 +01:00
City 23e59bbffe PixArt LoRA support 
Supports models from the example training code.
Weight loading will probably have to be changed for native formats.
 2023-12-22 16:16:14 +01:00
City a457da3f12 Add support for PixArt diffusers weights 2023-12-15 20:43:32 +01:00
City 80d5d9299b Temp fix for mask shape 2023-12-08 17:23:51 +01:00
City af62f98b4f PixArt KSampler support 
Thanks to @lawrence-cj for his help on this!
 2023-12-08 16:25:01 +01:00
City b31d8c517b PixArt small fix 
Issue #3 + some code changes from upstream.
 2023-12-06 21:31:07 +01:00
deinferno 0a176228f3 Fix PixArtT5TextEncode error 
'tuple' object has no attribute 'squeeze'
 2023-10-29 07:44:48 +05:00
City 5b42a090c4 xformers 2023-10-28 02:58:13 +02:00
City 814fe50852 PixArt support 2023-10-27 21:42:00 +02:00