Exempt IPEX from non_blocking previews fixing segmentation faults. #3708
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
chenbaiyujason
added a commit
to NeoWorldTeam/ComfyUI
that referenced
this pull request
Jun 19, 2024
commit 55f0dc1 Author: comfyanonymous <[email protected]> Date: Tue Jun 18 09:57:40 2024 -0400 Add soundfile dependency so that windows can save audio. commit a45df69 Author: comfyanonymous <[email protected]> Date: Mon Jun 17 22:48:23 2024 -0400 Basic tiled decoding for audio VAE. commit 379ff92 Author: Juanjuan <[email protected]> Date: Mon Jun 17 19:56:53 2024 +0800 fix app.js no graph defined (comfyanonymous#3754) * local test * fix "graph" not found * fix --------- Co-authored-by: Xiujuan Li <[email protected]> commit b7c473d Author: Janek Mann <[email protected]> Date: Mon Jun 17 12:55:06 2024 +0100 Fix lora keys for SimpleTuner (comfyanonymous#3759) commit 6425252 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:12:54 2024 -0400 Use fp16 as the default vae dtype for the audio VAE. commit 8ddc151 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:06:23 2024 -0400 Squash depreciation warning on new pytorch. commit ca9d300 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 11:47:32 2024 -0400 Better estimation for memory usage during audio VAE encoding/decoding. commit 746a041 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 03:10:04 2024 -0400 Fix VAEEncode with taesd3. commit 04e8798 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 02:04:24 2024 -0400 Improvements to the TAESD3 implementation. commit df7db0e Author: Dr.Lt.Data <[email protected]> Date: Sun Jun 16 15:03:53 2024 +0900 support TAESD3 (comfyanonymous#3738) commit bb1969c Author: comfyanonymous <[email protected]> Date: Sat Jun 15 12:14:56 2024 -0400 Initial support for the stable audio open model. commit 1281f93 Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:44:38 2024 -0400 Small optimization. commit f2e844e Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:25:03 2024 -0400 Optimize some unneeded if conditions in the sampling code. commit 0ec513d Author: comfyanonymous <[email protected]> Date: Sat Jun 15 01:08:12 2024 -0400 Add a --force-channels-last to inference models in channel last mode. commit 0e06b37 Author: comfyanonymous <[email protected]> Date: Fri Jun 14 18:18:53 2024 -0400 Print key names for easier debugging. commit 5eb98f0 Author: Simon Lui <[email protected]> Date: Thu Jun 13 15:51:14 2024 -0700 Exempt IPEX from non_blocking previews fixing segmentation faults. (comfyanonymous#3708) commit ac151ac Author: comfyanonymous <[email protected]> Date: Thu Jun 13 18:26:01 2024 -0400 Support SD3 diffusers lora. commit 37a08a4 Author: comfyanonymous <[email protected]> Date: Thu Jun 13 17:12:50 2024 -0400 Support setting weight offsets in weight patcher. commit 605e64f Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:39:33 2024 -0400 Fix lowvram issue. commit 0eaa34e Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:32:34 2024 -0400 Fix regular empty latent image not working with SD3 and custom sampler. commit 321e509 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 09:48:27 2024 -0400 Add link to SD3 example page to README. commit c8b5e08 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 02:24:39 2024 -0400 Default shift value on SD3 is 3.0 commit 1ddf512 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:07:58 2024 -0400 Don't auto convert clip and vae weights to fp16 when saving checkpoint. commit 32be358 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:02:07 2024 -0400 Save SD3 modelspec.architecture in CheckpointSave node. commit 694e0b4 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 00:49:00 2024 -0400 SD3 better memory usage estimation. commit 69c8d6d Author: comfyanonymous <[email protected]> Date: Tue Jun 11 23:27:39 2024 -0400 Single and dual clip loader nodes support SD3. You can use the CLIPLoader to use the t5xxl only or the DualCLIPLoader to use CLIP-L and CLIP-G only for sd3. commit 0e49211 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 17:03:26 2024 -0400 Load the SD3 T5xxl model in the same dtype stored in the checkpoint. commit 5889b7c Author: comfyanonymous <[email protected]> Date: Tue Jun 11 13:14:43 2024 -0400 Support multiple text encoder configurations on SD3. commit 1c34d33 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:37:22 2024 -0400 Update EmptySD3LatentImage to use 1024 resolution by default. commit 9424522 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:20:26 2024 -0400 Reuse code. commit 73ce178 Author: Dango233 <[email protected]> Date: Tue Jun 11 18:30:25 2024 +0800 Remove redundancy in mmdit.py (comfyanonymous#3685) commit 4134564 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 06:26:13 2024 -0400 Require safetensors library to be at least 0.4.2 for fp8 support.
csokun
pushed a commit
to csokun/ComfyUI
that referenced
this pull request
Jun 19, 2024
chenbaiyujason
added a commit
to NeoWorldTeam/ComfyUI
that referenced
this pull request
Jul 2, 2024
commit 2f03201 Author: comfyanonymous <[email protected]> Date: Tue Jul 2 01:32:23 2024 -0400 Remove some empty lines. commit 52aaee2 Author: shawnington <[email protected]> Date: Tue Jul 2 01:30:33 2024 -0400 Fix to comfyanonymous#3465. Prevent, resaving of duplicate images if overwrite not specified (comfyanonymous#3472) * Fix to comfyanonymous#3465. Prevent the, resaving of duplicate images if overwrite not specified This is a fix to comfyanonymous#3465 Adds function compare_image_hash to do a sha256 hash comparison between an uploaded image and existing images with matching file names. This changes the behavior so that only images having the same filename that are actually different are saved to input, existing images are instead now opened instead of resaved with increment. Currently, exact duplicates with the same filename are resave saved with an incremented filename in the format: <filename> (n).ext with the code: ``` while os.path.exists(filepath): filename = f"{split[0]} ({i}){split[1]}" filepath = os.path.join(full_output_folder, filename) i += 1 ``` This commit changes this to: ``` while os.path.exists(filepath): if compare_image_hash(filepath, image): image_is_duplicate = True break filename = f"{split[0]} ({i}){split[1]}" filepath = os.path.join(full_output_folder, filename) i += 1 ``` a check for if image_is_duplicate = False is done before saving the file. Currently, if you load the same image of a cat named cat.jpg into the LoadImage node 3 times, you will get 3 new files in your input folder with incremented file names. With this change, you will now only have the single copy of cat.jpg, that will be re-opened instead of re-saved. However if you load 3 different images of cats named cat.jpg, you will get the expected behavior of having: cat.jpg cat (1).jpg cat (2).jpg This saves space and clutter. After checking my own input folder, I have 800+ images that are duplicates that were resaved with incremented file names amounting to more than 5GB of duplicated data. * fixed typo in expression commit 1ef66b0 Author: Bob Du <[email protected]> Date: Tue Jul 2 06:02:42 2024 +0800 Add example for how to add custom API routes (comfyanonymous#3597) commit 9dd549e Author: Chenlei Hu <[email protected]> Date: Mon Jul 1 17:54:03 2024 -0400 Add `--no-custom-node` cmd flag (comfyanonymous#3903) * Add --no-custom-node cmd flag * nit commit b82d67d Author: Peter Crabtree <[email protected]> Date: Mon Jul 1 17:42:17 2024 -0400 Add SamplerEulerAncestralCFG++ custom sampler node (comfyanonymous#3901) (for eta and s_noise) commit 755c48d Author: Hayden Reeve <[email protected]> Date: Tue Jul 2 05:21:12 2024 +0800 Fix several typos in example_node.py.example (comfyanonymous#3204) This change includes corrections for several spelling errors in the documentation of example_node.py.example file. These were previously raised by comfyanonymous#3157, but they missed a few. commit 5dccfef Author: comfyanonymous <[email protected]> Date: Mon Jul 1 17:17:25 2024 -0400 Switch nightly pytorch standalone package to lzma2. commit 0cd4a6a Author: YAN Wenkun <[email protected]> Date: Tue Jul 2 05:15:49 2024 +0800 Fine-tuning GitHub Actions (comfyanonymous#3169) * Bumping GitHub Actions versions * Using LZMA2 for 7zip compression in Windows packaging commit 601b4b6 Author: Robin Huang <[email protected]> Date: Mon Jul 1 10:51:00 2024 -0700 Add CONTRIBUTING.md (comfyanonymous#3910) * Create CONTRIBUTING.md * Add feature-request channel link. * Remove discord links for channels. commit e53b159 Author: ruucm <[email protected]> Date: Mon Jul 1 10:45:34 2024 -0700 enable cmd shortcuts for mac (mute & bypass) (comfyanonymous#3792) commit 7c5fa7f Author: Chenlei Hu <[email protected]> Date: Mon Jul 1 12:10:44 2024 -0400 Fix loadGraphData func call (comfyanonymous#3918) commit 521421f Author: comfyanonymous <[email protected]> Date: Sun Jun 30 15:51:54 2024 -0400 Fix workflow not importing from flac files on some systems. commit dbb7dd3 Author: comfyanonymous <[email protected]> Date: Sun Jun 30 00:15:49 2024 -0400 Add to readme that Stable Audio is supported. commit 05e8316 Author: comfyanonymous <[email protected]> Date: Sat Jun 29 11:59:48 2024 -0400 Switch to the real cfg++ method in the samplers. The old _pp ones will be updated automatically to the regular ones with 2x the cfg. My fault for not checking what the "_pp" samplers actually did. commit fbb7a1f Author: comfyanonymous <[email protected]> Date: Sat Jun 29 01:33:22 2024 -0400 PreviewAudio node. commit c39cf7f Author: Robin Huang <[email protected]> Date: Fri Jun 28 13:09:55 2024 -0700 Revert "Add integration test for Linux with Nvidia GPU. comfyanonymous#3884 (comfyanonymous#3895)" (comfyanonymous#3905) This reverts commit 449bf52. commit 02cac1d Author: Robin Huang <[email protected]> Date: Fri Jun 28 13:09:39 2024 -0700 Revert "Add macOs integration test for default workflow. (comfyanonymous#3898)" (comfyanonymous#3904) This reverts commit 97b409c. commit 7ecb2ec Author: comfyanonymous <[email protected]> Date: Fri Jun 28 02:55:36 2024 -0400 Audio second setting in EmptyLatentAudio. commit 0d9009c Author: pythongosssss <[email protected]> Date: Fri Jun 28 06:07:19 2024 +0100 New menu/workflows fixes (comfyanonymous#3900) * Fix auto queue * Detect added nodes via search * Fix loading workflows * Add button click style commit 264caca Author: comfyanonymous <[email protected]> Date: Wed Jun 26 16:14:47 2024 -0400 ControlNetApplySD3 node can now be used to use SD3 controlnets. commit f8f7568 Author: comfyanonymous <[email protected]> Date: Tue Jun 25 23:40:44 2024 -0400 Basic SD3 controlnet implementation. Still missing the node to properly use it. commit 66aaa14 Author: comfyanonymous <[email protected]> Date: Tue Jun 25 17:02:05 2024 -0400 Controlnet refactor. commit 97b409c Author: Robin Huang <[email protected]> Date: Thu Jun 27 13:10:16 2024 -0700 Add macOs integration test for default workflow. (comfyanonymous#3898) commit 449bf52 Author: Robin Huang <[email protected]> Date: Thu Jun 27 13:08:26 2024 -0700 Add integration test for Linux with Nvidia GPU. comfyanonymous#3884 (comfyanonymous#3895) * Add linux integration test. * Fix directory path. * Add paths ignore. * Fix conda env directory path. commit 8ceb5a0 Author: comfyanonymous <[email protected]> Date: Thu Jun 27 11:06:52 2024 -0400 Support saving stable audio checkpoint that can be loaded back. commit 5ff3d4e Author: Chenlei Hu <[email protected]> Date: Thu Jun 27 09:13:52 2024 -0400 Fix audio upload when no audio in input dir (comfyanonymous#3891) commit 4f9d2b0 Author: comfyanonymous <[email protected]> Date: Thu Jun 27 02:54:15 2024 -0400 Remove print. commit 4650e7d Author: comfyanonymous <[email protected]> Date: Thu Jun 27 01:28:59 2024 -0400 Save and load workflow from the flac files output by SaveAudio. commit 3b423af Author: Chenlei Hu <[email protected]> Date: Thu Jun 27 00:22:55 2024 -0400 Add audio widget (comfyanonymous#3863) * Add audio widget * Fix audio bugs * Add CSS * Populate audio widget when load history commit 44947e7 Author: comfyanonymous <[email protected]> Date: Wed Jun 26 22:40:05 2024 -0400 Add DEIS order 3 sampler. Order 4 seems to give bad results. commit 175fe02 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:59:19 2024 -0400 Ignore .vscode/ (comfyanonymous#3879) commit bc5a0f1 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:59:09 2024 -0400 Ignore *.log (comfyanonymous#3880) commit a3e83f6 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:58:56 2024 -0400 Update test ref (comfyanonymous#3882) * Update ref * Disable some tests commit f12fa1d Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 09:09:21 2024 -0400 Enable browser tests on push (comfyanonymous#3878) commit e3579f3 Author: pythongosssss <[email protected]> Date: Wed Jun 26 14:08:48 2024 +0100 Fix merge issue breaking api json loading (comfyanonymous#3876) commit edfce78 Author: Alex "mcmonkey" Goodwin <[email protected]> Date: Tue Jun 25 22:37:27 2024 -0700 add issue templates for ComfyUI Issues Page (comfyanonymous#3868) commit e99d97a Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 01:23:55 2024 -0400 Remove duplicated Reset View button (comfyanonymous#3865) * Remove duplicated Reset View button * Disable flaky test commit 69d710e Author: comfyanonymous <[email protected]> Date: Tue Jun 25 07:41:52 2024 -0400 Implement my alternative take on CFG++ as the euler_pp sampler. Add euler_ancestral_pp which is the ancestral version of euler with the same modification. commit 90aebb6 Author: pythongosssss <[email protected]> Date: Tue Jun 25 11:49:25 2024 +0100 New Menu & Workflow Management (comfyanonymous#3112) * menu * wip * wip * wip * wip * wip * workflow saving/loading * Support inserting workflows Move buttosn to top of lists * fix session storage implement renaming * temp * refactor, better workflow instance management * wip * progress on progress * added send to workflow various fixes * Support multiple image loaders * Support dynamic size breakpoints based on content * various fixes add close unsaved warning * Add filtering tree * prevent renaming unsaved * fix zindex on hover * fix top offset * use filename as workflow name * resize on setting change * hide element until it is drawn * remove glow * Fix export name * Fix test, revert accidental changes to groupNode * Fix colors on all themes * show hover items on smaller screen (mobile) * remove debugging code * dialog fix * Dont reorder open workflows Allow elements around canvas * Toggle body display on setting change * Fix menu disappearing on chrome * Increase delay when typing, remove margin on Safari, fix dialog location * Fix overflow issue on iOS * Add reset view button Prevent view changes causing history entries * Bottom menu wip * Various fixes * Fix merge * Fix breaking old menu position * Fix merge adding restore view to loadGraphData commit eab211b Author: comfyanonymous <[email protected]> Date: Mon Jun 24 16:55:20 2024 -0400 Resample audio to 44100 when VAE encoding it. commit 866f54d Author: Chenlei Hu <[email protected]> Date: Mon Jun 24 14:47:28 2024 -0400 Add browser test action synced with TS repo (comfyanonymous#3852) * Add browser test action * Add npm install task commit 73ca780 Author: comfyanonymous <[email protected]> Date: Sun Jun 23 13:21:18 2024 -0400 Add SamplerEulerCFG++ node. This node should match the DDIM implementation of CFG++ when "regular" is selected. "alternative" is a slightly different take on CFG++ commit 2f360ae Author: comfyanonymous <[email protected]> Date: Sat Jun 22 13:08:04 2024 -0400 Support OneTrainer SD3 lora format. commit 4ef1479 Author: comfyanonymous <[email protected]> Date: Sat Jun 22 11:45:58 2024 -0400 Multi dimension tiled scale function and tiled VAE audio encoding fallback. commit 887a634 Author: comfyanonymous <[email protected]> Date: Fri Jun 21 08:41:31 2024 -0400 Proper ModelMergeSD3_2B node. commit 1e2839f Author: comfyanonymous <[email protected]> Date: Thu Jun 20 16:50:31 2024 -0400 More proper tiled audio decoding. commit d5efde8 Author: comfyanonymous <[email protected]> Date: Thu Jun 20 08:51:49 2024 -0400 Add ipndm_v sampler, works best with the exponential scheduler. commit 45e10ca Author: Zhenyu Zhou <[email protected]> Date: Thu Jun 20 20:12:15 2024 +0800 feat: add gits scheduler (comfyanonymous#3769) commit d7f0964 Author: Chenlei Hu <[email protected]> Date: Wed Jun 19 22:36:31 2024 -0400 Fix routes (comfyanonymous#3790) commit 028a583 Author: comfyanonymous <[email protected]> Date: Wed Jun 19 22:32:04 2024 -0400 Fix issue with full diffusers SD3 loras. commit 0d6a579 Author: comfyanonymous <[email protected]> Date: Wed Jun 19 21:46:37 2024 -0400 Support loading diffusers SD3 model format with UNETLoader node. commit b08a9dd Author: comfyanonymous <[email protected]> Date: Wed Jun 19 20:20:35 2024 -0400 Remove empty line. commit eee815e Author: Mario Klingemann <[email protected]> Date: Wed Jun 19 22:42:41 2024 +0200 Update sd1_clip.py (comfyanonymous#3684) Made token instance check more flexible so it also works with integers from numpy arrays or long tensors commit e11052a Author: comfyanonymous <[email protected]> Date: Wed Jun 19 16:32:30 2024 -0400 Add ipndm sampler. commit 97ae6ef Author: Chenlei Hu <[email protected]> Date: Wed Jun 19 10:39:17 2024 -0400 Add api/ prefix to api endpoints (comfyanonymous#3779) commit 3914d5a Author: comfyanonymous <[email protected]> Date: Wed Jun 19 10:01:43 2024 -0400 Support full SD3 loras. commit 55f0dc1 Author: comfyanonymous <[email protected]> Date: Tue Jun 18 09:57:40 2024 -0400 Add soundfile dependency so that windows can save audio. commit a45df69 Author: comfyanonymous <[email protected]> Date: Mon Jun 17 22:48:23 2024 -0400 Basic tiled decoding for audio VAE. commit 379ff92 Author: Juanjuan <[email protected]> Date: Mon Jun 17 19:56:53 2024 +0800 fix app.js no graph defined (comfyanonymous#3754) * local test * fix "graph" not found * fix --------- Co-authored-by: Xiujuan Li <[email protected]> commit b7c473d Author: Janek Mann <[email protected]> Date: Mon Jun 17 12:55:06 2024 +0100 Fix lora keys for SimpleTuner (comfyanonymous#3759) commit 6425252 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:12:54 2024 -0400 Use fp16 as the default vae dtype for the audio VAE. commit 8ddc151 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:06:23 2024 -0400 Squash depreciation warning on new pytorch. commit ca9d300 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 11:47:32 2024 -0400 Better estimation for memory usage during audio VAE encoding/decoding. commit 746a041 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 03:10:04 2024 -0400 Fix VAEEncode with taesd3. commit 04e8798 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 02:04:24 2024 -0400 Improvements to the TAESD3 implementation. commit df7db0e Author: Dr.Lt.Data <[email protected]> Date: Sun Jun 16 15:03:53 2024 +0900 support TAESD3 (comfyanonymous#3738) commit bb1969c Author: comfyanonymous <[email protected]> Date: Sat Jun 15 12:14:56 2024 -0400 Initial support for the stable audio open model. commit 1281f93 Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:44:38 2024 -0400 Small optimization. commit f2e844e Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:25:03 2024 -0400 Optimize some unneeded if conditions in the sampling code. commit 0ec513d Author: comfyanonymous <[email protected]> Date: Sat Jun 15 01:08:12 2024 -0400 Add a --force-channels-last to inference models in channel last mode. commit 0e06b37 Author: comfyanonymous <[email protected]> Date: Fri Jun 14 18:18:53 2024 -0400 Print key names for easier debugging. commit 5eb98f0 Author: Simon Lui <[email protected]> Date: Thu Jun 13 15:51:14 2024 -0700 Exempt IPEX from non_blocking previews fixing segmentation faults. (comfyanonymous#3708) commit ac151ac Author: comfyanonymous <[email protected]> Date: Thu Jun 13 18:26:01 2024 -0400 Support SD3 diffusers lora. commit 37a08a4 Author: comfyanonymous <[email protected]> Date: Thu Jun 13 17:12:50 2024 -0400 Support setting weight offsets in weight patcher. commit 605e64f Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:39:33 2024 -0400 Fix lowvram issue. commit 0eaa34e Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:32:34 2024 -0400 Fix regular empty latent image not working with SD3 and custom sampler. commit 321e509 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 09:48:27 2024 -0400 Add link to SD3 example page to README. commit c8b5e08 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 02:24:39 2024 -0400 Default shift value on SD3 is 3.0 commit 1ddf512 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:07:58 2024 -0400 Don't auto convert clip and vae weights to fp16 when saving checkpoint. commit 32be358 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:02:07 2024 -0400 Save SD3 modelspec.architecture in CheckpointSave node. commit 694e0b4 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 00:49:00 2024 -0400 SD3 better memory usage estimation. commit 69c8d6d Author: comfyanonymous <[email protected]> Date: Tue Jun 11 23:27:39 2024 -0400 Single and dual clip loader nodes support SD3. You can use the CLIPLoader to use the t5xxl only or the DualCLIPLoader to use CLIP-L and CLIP-G only for sd3. commit 0e49211 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 17:03:26 2024 -0400 Load the SD3 T5xxl model in the same dtype stored in the checkpoint. commit 5889b7c Author: comfyanonymous <[email protected]> Date: Tue Jun 11 13:14:43 2024 -0400 Support multiple text encoder configurations on SD3. commit 1c34d33 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:37:22 2024 -0400 Update EmptySD3LatentImage to use 1024 resolution by default. commit 9424522 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:20:26 2024 -0400 Reuse code. commit 73ce178 Author: Dango233 <[email protected]> Date: Tue Jun 11 18:30:25 2024 +0800 Remove redundancy in mmdit.py (comfyanonymous#3685) commit 4134564 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 06:26:13 2024 -0400 Require safetensors library to be at least 0.4.2 for fp8 support. # Conflicts: # .gitignore # README.md # comfy/ldm/modules/diffusionmodules/mmdit.py # comfy/lora.py # comfy/model_base.py resolved by upstream/master version # comfy/model_patcher.py resolved by upstream/master version # comfy/sd.py resolved by upstream/master version # comfy/supported_models.py resolved by upstream/master version # comfy_extras/nodes_audio.py resolved by upstream/master version
chenbaiyujason
added a commit
to NeoWorldTeam/ComfyUI
that referenced
this pull request
Jul 3, 2024
commit 537f35c Author: comfyanonymous <[email protected]> Date: Tue Jul 2 20:21:51 2024 -0400 Don't update dict if contiguous. commit 3f46362 Author: Alex "mcmonkey" Goodwin <[email protected]> Date: Tue Jul 2 17:16:33 2024 -0700 fix non-contiguous tensor saving (from channels-last) (comfyanonymous#3932) commit 01991f7 Author: comfyanonymous <[email protected]> Date: Tue Jul 2 12:21:08 2024 -0400 Fix SamplerEulerCFGpp node. commit 2f03201 Author: comfyanonymous <[email protected]> Date: Tue Jul 2 01:32:23 2024 -0400 Remove some empty lines. commit 52aaee2 Author: shawnington <[email protected]> Date: Tue Jul 2 01:30:33 2024 -0400 Fix to comfyanonymous#3465. Prevent, resaving of duplicate images if overwrite not specified (comfyanonymous#3472) * Fix to comfyanonymous#3465. Prevent the, resaving of duplicate images if overwrite not specified This is a fix to comfyanonymous#3465 Adds function compare_image_hash to do a sha256 hash comparison between an uploaded image and existing images with matching file names. This changes the behavior so that only images having the same filename that are actually different are saved to input, existing images are instead now opened instead of resaved with increment. Currently, exact duplicates with the same filename are resave saved with an incremented filename in the format: <filename> (n).ext with the code: ``` while os.path.exists(filepath): filename = f"{split[0]} ({i}){split[1]}" filepath = os.path.join(full_output_folder, filename) i += 1 ``` This commit changes this to: ``` while os.path.exists(filepath): if compare_image_hash(filepath, image): image_is_duplicate = True break filename = f"{split[0]} ({i}){split[1]}" filepath = os.path.join(full_output_folder, filename) i += 1 ``` a check for if image_is_duplicate = False is done before saving the file. Currently, if you load the same image of a cat named cat.jpg into the LoadImage node 3 times, you will get 3 new files in your input folder with incremented file names. With this change, you will now only have the single copy of cat.jpg, that will be re-opened instead of re-saved. However if you load 3 different images of cats named cat.jpg, you will get the expected behavior of having: cat.jpg cat (1).jpg cat (2).jpg This saves space and clutter. After checking my own input folder, I have 800+ images that are duplicates that were resaved with incremented file names amounting to more than 5GB of duplicated data. * fixed typo in expression commit 1ef66b0 Author: Bob Du <[email protected]> Date: Tue Jul 2 06:02:42 2024 +0800 Add example for how to add custom API routes (comfyanonymous#3597) commit 9dd549e Author: Chenlei Hu <[email protected]> Date: Mon Jul 1 17:54:03 2024 -0400 Add `--no-custom-node` cmd flag (comfyanonymous#3903) * Add --no-custom-node cmd flag * nit commit b82d67d Author: Peter Crabtree <[email protected]> Date: Mon Jul 1 17:42:17 2024 -0400 Add SamplerEulerAncestralCFG++ custom sampler node (comfyanonymous#3901) (for eta and s_noise) commit 755c48d Author: Hayden Reeve <[email protected]> Date: Tue Jul 2 05:21:12 2024 +0800 Fix several typos in example_node.py.example (comfyanonymous#3204) This change includes corrections for several spelling errors in the documentation of example_node.py.example file. These were previously raised by comfyanonymous#3157, but they missed a few. commit 5dccfef Author: comfyanonymous <[email protected]> Date: Mon Jul 1 17:17:25 2024 -0400 Switch nightly pytorch standalone package to lzma2. commit 0cd4a6a Author: YAN Wenkun <[email protected]> Date: Tue Jul 2 05:15:49 2024 +0800 Fine-tuning GitHub Actions (comfyanonymous#3169) * Bumping GitHub Actions versions * Using LZMA2 for 7zip compression in Windows packaging commit 601b4b6 Author: Robin Huang <[email protected]> Date: Mon Jul 1 10:51:00 2024 -0700 Add CONTRIBUTING.md (comfyanonymous#3910) * Create CONTRIBUTING.md * Add feature-request channel link. * Remove discord links for channels. commit e53b159 Author: ruucm <[email protected]> Date: Mon Jul 1 10:45:34 2024 -0700 enable cmd shortcuts for mac (mute & bypass) (comfyanonymous#3792) commit 7c5fa7f Author: Chenlei Hu <[email protected]> Date: Mon Jul 1 12:10:44 2024 -0400 Fix loadGraphData func call (comfyanonymous#3918) commit 521421f Author: comfyanonymous <[email protected]> Date: Sun Jun 30 15:51:54 2024 -0400 Fix workflow not importing from flac files on some systems. commit dbb7dd3 Author: comfyanonymous <[email protected]> Date: Sun Jun 30 00:15:49 2024 -0400 Add to readme that Stable Audio is supported. commit 05e8316 Author: comfyanonymous <[email protected]> Date: Sat Jun 29 11:59:48 2024 -0400 Switch to the real cfg++ method in the samplers. The old _pp ones will be updated automatically to the regular ones with 2x the cfg. My fault for not checking what the "_pp" samplers actually did. commit fbb7a1f Author: comfyanonymous <[email protected]> Date: Sat Jun 29 01:33:22 2024 -0400 PreviewAudio node. commit c39cf7f Author: Robin Huang <[email protected]> Date: Fri Jun 28 13:09:55 2024 -0700 Revert "Add integration test for Linux with Nvidia GPU. comfyanonymous#3884 (comfyanonymous#3895)" (comfyanonymous#3905) This reverts commit 449bf52. commit 02cac1d Author: Robin Huang <[email protected]> Date: Fri Jun 28 13:09:39 2024 -0700 Revert "Add macOs integration test for default workflow. (comfyanonymous#3898)" (comfyanonymous#3904) This reverts commit 97b409c. commit 7ecb2ec Author: comfyanonymous <[email protected]> Date: Fri Jun 28 02:55:36 2024 -0400 Audio second setting in EmptyLatentAudio. commit 0d9009c Author: pythongosssss <[email protected]> Date: Fri Jun 28 06:07:19 2024 +0100 New menu/workflows fixes (comfyanonymous#3900) * Fix auto queue * Detect added nodes via search * Fix loading workflows * Add button click style commit 264caca Author: comfyanonymous <[email protected]> Date: Wed Jun 26 16:14:47 2024 -0400 ControlNetApplySD3 node can now be used to use SD3 controlnets. commit f8f7568 Author: comfyanonymous <[email protected]> Date: Tue Jun 25 23:40:44 2024 -0400 Basic SD3 controlnet implementation. Still missing the node to properly use it. commit 66aaa14 Author: comfyanonymous <[email protected]> Date: Tue Jun 25 17:02:05 2024 -0400 Controlnet refactor. commit 97b409c Author: Robin Huang <[email protected]> Date: Thu Jun 27 13:10:16 2024 -0700 Add macOs integration test for default workflow. (comfyanonymous#3898) commit 449bf52 Author: Robin Huang <[email protected]> Date: Thu Jun 27 13:08:26 2024 -0700 Add integration test for Linux with Nvidia GPU. comfyanonymous#3884 (comfyanonymous#3895) * Add linux integration test. * Fix directory path. * Add paths ignore. * Fix conda env directory path. commit 8ceb5a0 Author: comfyanonymous <[email protected]> Date: Thu Jun 27 11:06:52 2024 -0400 Support saving stable audio checkpoint that can be loaded back. commit 5ff3d4e Author: Chenlei Hu <[email protected]> Date: Thu Jun 27 09:13:52 2024 -0400 Fix audio upload when no audio in input dir (comfyanonymous#3891) commit 4f9d2b0 Author: comfyanonymous <[email protected]> Date: Thu Jun 27 02:54:15 2024 -0400 Remove print. commit 4650e7d Author: comfyanonymous <[email protected]> Date: Thu Jun 27 01:28:59 2024 -0400 Save and load workflow from the flac files output by SaveAudio. commit 3b423af Author: Chenlei Hu <[email protected]> Date: Thu Jun 27 00:22:55 2024 -0400 Add audio widget (comfyanonymous#3863) * Add audio widget * Fix audio bugs * Add CSS * Populate audio widget when load history commit 44947e7 Author: comfyanonymous <[email protected]> Date: Wed Jun 26 22:40:05 2024 -0400 Add DEIS order 3 sampler. Order 4 seems to give bad results. commit 175fe02 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:59:19 2024 -0400 Ignore .vscode/ (comfyanonymous#3879) commit bc5a0f1 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:59:09 2024 -0400 Ignore *.log (comfyanonymous#3880) commit a3e83f6 Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 19:58:56 2024 -0400 Update test ref (comfyanonymous#3882) * Update ref * Disable some tests commit f12fa1d Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 09:09:21 2024 -0400 Enable browser tests on push (comfyanonymous#3878) commit e3579f3 Author: pythongosssss <[email protected]> Date: Wed Jun 26 14:08:48 2024 +0100 Fix merge issue breaking api json loading (comfyanonymous#3876) commit edfce78 Author: Alex "mcmonkey" Goodwin <[email protected]> Date: Tue Jun 25 22:37:27 2024 -0700 add issue templates for ComfyUI Issues Page (comfyanonymous#3868) commit e99d97a Author: Chenlei Hu <[email protected]> Date: Wed Jun 26 01:23:55 2024 -0400 Remove duplicated Reset View button (comfyanonymous#3865) * Remove duplicated Reset View button * Disable flaky test commit 69d710e Author: comfyanonymous <[email protected]> Date: Tue Jun 25 07:41:52 2024 -0400 Implement my alternative take on CFG++ as the euler_pp sampler. Add euler_ancestral_pp which is the ancestral version of euler with the same modification. commit 90aebb6 Author: pythongosssss <[email protected]> Date: Tue Jun 25 11:49:25 2024 +0100 New Menu & Workflow Management (comfyanonymous#3112) * menu * wip * wip * wip * wip * wip * workflow saving/loading * Support inserting workflows Move buttosn to top of lists * fix session storage implement renaming * temp * refactor, better workflow instance management * wip * progress on progress * added send to workflow various fixes * Support multiple image loaders * Support dynamic size breakpoints based on content * various fixes add close unsaved warning * Add filtering tree * prevent renaming unsaved * fix zindex on hover * fix top offset * use filename as workflow name * resize on setting change * hide element until it is drawn * remove glow * Fix export name * Fix test, revert accidental changes to groupNode * Fix colors on all themes * show hover items on smaller screen (mobile) * remove debugging code * dialog fix * Dont reorder open workflows Allow elements around canvas * Toggle body display on setting change * Fix menu disappearing on chrome * Increase delay when typing, remove margin on Safari, fix dialog location * Fix overflow issue on iOS * Add reset view button Prevent view changes causing history entries * Bottom menu wip * Various fixes * Fix merge * Fix breaking old menu position * Fix merge adding restore view to loadGraphData commit eab211b Author: comfyanonymous <[email protected]> Date: Mon Jun 24 16:55:20 2024 -0400 Resample audio to 44100 when VAE encoding it. commit 866f54d Author: Chenlei Hu <[email protected]> Date: Mon Jun 24 14:47:28 2024 -0400 Add browser test action synced with TS repo (comfyanonymous#3852) * Add browser test action * Add npm install task commit 73ca780 Author: comfyanonymous <[email protected]> Date: Sun Jun 23 13:21:18 2024 -0400 Add SamplerEulerCFG++ node. This node should match the DDIM implementation of CFG++ when "regular" is selected. "alternative" is a slightly different take on CFG++ commit 2f360ae Author: comfyanonymous <[email protected]> Date: Sat Jun 22 13:08:04 2024 -0400 Support OneTrainer SD3 lora format. commit 4ef1479 Author: comfyanonymous <[email protected]> Date: Sat Jun 22 11:45:58 2024 -0400 Multi dimension tiled scale function and tiled VAE audio encoding fallback. commit 887a634 Author: comfyanonymous <[email protected]> Date: Fri Jun 21 08:41:31 2024 -0400 Proper ModelMergeSD3_2B node. commit 1e2839f Author: comfyanonymous <[email protected]> Date: Thu Jun 20 16:50:31 2024 -0400 More proper tiled audio decoding. commit d5efde8 Author: comfyanonymous <[email protected]> Date: Thu Jun 20 08:51:49 2024 -0400 Add ipndm_v sampler, works best with the exponential scheduler. commit 45e10ca Author: Zhenyu Zhou <[email protected]> Date: Thu Jun 20 20:12:15 2024 +0800 feat: add gits scheduler (comfyanonymous#3769) commit d7f0964 Author: Chenlei Hu <[email protected]> Date: Wed Jun 19 22:36:31 2024 -0400 Fix routes (comfyanonymous#3790) commit 028a583 Author: comfyanonymous <[email protected]> Date: Wed Jun 19 22:32:04 2024 -0400 Fix issue with full diffusers SD3 loras. commit 0d6a579 Author: comfyanonymous <[email protected]> Date: Wed Jun 19 21:46:37 2024 -0400 Support loading diffusers SD3 model format with UNETLoader node. commit b08a9dd Author: comfyanonymous <[email protected]> Date: Wed Jun 19 20:20:35 2024 -0400 Remove empty line. commit eee815e Author: Mario Klingemann <[email protected]> Date: Wed Jun 19 22:42:41 2024 +0200 Update sd1_clip.py (comfyanonymous#3684) Made token instance check more flexible so it also works with integers from numpy arrays or long tensors commit e11052a Author: comfyanonymous <[email protected]> Date: Wed Jun 19 16:32:30 2024 -0400 Add ipndm sampler. commit 97ae6ef Author: Chenlei Hu <[email protected]> Date: Wed Jun 19 10:39:17 2024 -0400 Add api/ prefix to api endpoints (comfyanonymous#3779) commit 3914d5a Author: comfyanonymous <[email protected]> Date: Wed Jun 19 10:01:43 2024 -0400 Support full SD3 loras. commit 55f0dc1 Author: comfyanonymous <[email protected]> Date: Tue Jun 18 09:57:40 2024 -0400 Add soundfile dependency so that windows can save audio. commit a45df69 Author: comfyanonymous <[email protected]> Date: Mon Jun 17 22:48:23 2024 -0400 Basic tiled decoding for audio VAE. commit 379ff92 Author: Juanjuan <[email protected]> Date: Mon Jun 17 19:56:53 2024 +0800 fix app.js no graph defined (comfyanonymous#3754) * local test * fix "graph" not found * fix --------- Co-authored-by: Xiujuan Li <[email protected]> commit b7c473d Author: Janek Mann <[email protected]> Date: Mon Jun 17 12:55:06 2024 +0100 Fix lora keys for SimpleTuner (comfyanonymous#3759) commit 6425252 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:12:54 2024 -0400 Use fp16 as the default vae dtype for the audio VAE. commit 8ddc151 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 13:06:23 2024 -0400 Squash depreciation warning on new pytorch. commit ca9d300 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 11:47:32 2024 -0400 Better estimation for memory usage during audio VAE encoding/decoding. commit 746a041 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 03:10:04 2024 -0400 Fix VAEEncode with taesd3. commit 04e8798 Author: comfyanonymous <[email protected]> Date: Sun Jun 16 02:04:24 2024 -0400 Improvements to the TAESD3 implementation. commit df7db0e Author: Dr.Lt.Data <[email protected]> Date: Sun Jun 16 15:03:53 2024 +0900 support TAESD3 (comfyanonymous#3738) commit bb1969c Author: comfyanonymous <[email protected]> Date: Sat Jun 15 12:14:56 2024 -0400 Initial support for the stable audio open model. commit 1281f93 Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:44:38 2024 -0400 Small optimization. commit f2e844e Author: comfyanonymous <[email protected]> Date: Sat Jun 15 02:25:03 2024 -0400 Optimize some unneeded if conditions in the sampling code. commit 0ec513d Author: comfyanonymous <[email protected]> Date: Sat Jun 15 01:08:12 2024 -0400 Add a --force-channels-last to inference models in channel last mode. commit 0e06b37 Author: comfyanonymous <[email protected]> Date: Fri Jun 14 18:18:53 2024 -0400 Print key names for easier debugging. commit 5eb98f0 Author: Simon Lui <[email protected]> Date: Thu Jun 13 15:51:14 2024 -0700 Exempt IPEX from non_blocking previews fixing segmentation faults. (comfyanonymous#3708) commit ac151ac Author: comfyanonymous <[email protected]> Date: Thu Jun 13 18:26:01 2024 -0400 Support SD3 diffusers lora. commit 37a08a4 Author: comfyanonymous <[email protected]> Date: Thu Jun 13 17:12:50 2024 -0400 Support setting weight offsets in weight patcher. commit 605e64f Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:39:33 2024 -0400 Fix lowvram issue. commit 0eaa34e Author: comfyanonymous <[email protected]> Date: Wed Jun 12 10:32:34 2024 -0400 Fix regular empty latent image not working with SD3 and custom sampler. commit 321e509 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 09:48:27 2024 -0400 Add link to SD3 example page to README. commit c8b5e08 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 02:24:39 2024 -0400 Default shift value on SD3 is 3.0 commit 1ddf512 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:07:58 2024 -0400 Don't auto convert clip and vae weights to fp16 when saving checkpoint. commit 32be358 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 01:02:07 2024 -0400 Save SD3 modelspec.architecture in CheckpointSave node. commit 694e0b4 Author: comfyanonymous <[email protected]> Date: Wed Jun 12 00:49:00 2024 -0400 SD3 better memory usage estimation. commit 69c8d6d Author: comfyanonymous <[email protected]> Date: Tue Jun 11 23:27:39 2024 -0400 Single and dual clip loader nodes support SD3. You can use the CLIPLoader to use the t5xxl only or the DualCLIPLoader to use CLIP-L and CLIP-G only for sd3. commit 0e49211 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 17:03:26 2024 -0400 Load the SD3 T5xxl model in the same dtype stored in the checkpoint. commit 5889b7c Author: comfyanonymous <[email protected]> Date: Tue Jun 11 13:14:43 2024 -0400 Support multiple text encoder configurations on SD3. commit 1c34d33 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:37:22 2024 -0400 Update EmptySD3LatentImage to use 1024 resolution by default. commit 9424522 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 07:20:26 2024 -0400 Reuse code. commit 73ce178 Author: Dango233 <[email protected]> Date: Tue Jun 11 18:30:25 2024 +0800 Remove redundancy in mmdit.py (comfyanonymous#3685) commit 4134564 Author: comfyanonymous <[email protected]> Date: Tue Jun 11 06:26:13 2024 -0400 Require safetensors library to be at least 0.4.2 for fp8 support. # Conflicts: # .gitignore # comfy_extras/nodes_advanced_samplers.py resolved by upstream/master version
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Sorry, forgot to send this patch earlier. See intel/intel-extension-for-pytorch#636 (comment) and the thread in general for the discussion on this but it seems like the argument is that the runtime is fine and that the XPU-CPU transfer is not ready yet in IPEX before returning the image data so it needs to be set to non-blocking.