4 Pull requests merged by 4 people

• fix app.js no graph defined

  #3754 merged Jun 17, 2024

• Fix lora keys for SimpleTuner

  #3759 merged Jun 17, 2024

• feat: support TAESD3

  #3738 merged Jun 16, 2024

• Exempt IPEX from non_blocking previews fixing segmentation faults.

  #3708 merged Jun 13, 2024

4 Pull requests opened by 4 people

• feat: support node smart offload, reduce peak VRAM/RAM usage

  #3709 opened Jun 13, 2024

• 18x faster！Update nodes_mask.py

  #3727 opened Jun 14, 2024

• DOC: Add custom_nodes to the example extra_model_paths.

  #3744 opened Jun 16, 2024

• feat: add gits scheduler

  #3769 opened Jun 18, 2024

30 Issues closed by 15 people

• Unable to access port 8188(无法访问 8188 端口)

  #3750 closed Jun 17, 2024

• A problem while trying to run the basic way

  #3733 closed Jun 16, 2024

• source code string cannot contain null bytes

  #3710 closed Jun 15, 2024

• How to show each image AS it is generated, not all at the end?

  #3711 closed Jun 15, 2024

• After the upgrade, the node cannot be converted to input.

  #3712 closed Jun 15, 2024

• Required input is missing: upscale_model_name

  #3713 closed Jun 15, 2024

• [ONNXRuntimeError] ComfyUI-WD14-Tagger

  #3716 closed Jun 15, 2024

• Is it possible to train LoRa on model merged by ModelMergeSimple?

  #3718 closed Jun 15, 2024

• too many values to unpack (expected 3)

  #3719 closed Jun 15, 2024

• How to solve it? Can anyone help?

  #3726 closed Jun 15, 2024

• [BUG][37a08a4] using fooocus occuer Error

  #3730 closed Jun 15, 2024

• Execution queue error

  #3723 closed Jun 15, 2024

• RuntimeError: CUDA error: HIPBLAS_STATUS_NOT_SUPPORTED when calling `HIPBLAS_STATUS_NOT_SUPPORTED`

  #3698 closed Jun 14, 2024

• Stuck for inpaiting

  #3724 closed Jun 14, 2024

• 【ipex】Unexpected Keyword Argument 'concat_linear' Error in CheckpointLoaderSimple Execution

  #3579 closed Jun 14, 2024

• error: Your local changes to the following files would be overwritten by merge:

  #3700 closed Jun 14, 2024

• Exception during processing!!! GET was unable to find an engine to execute this computation

  #3705 closed Jun 13, 2024

• Import/Ksampler Issue

  #3707 closed Jun 13, 2024

• After the update, the Ksample is no longer possible to use condition combine with set mask...The last available version is 742d572

  #3704 closed Jun 13, 2024

• SD3 Image resolution problem

  #3697 closed Jun 13, 2024

• update or rather sabotage...

  #3706 closed Jun 13, 2024

• Error when loading sd3_medium_incl_clips_t5xxlfp8.safetensors

  #3696 closed Jun 13, 2024

• Stable Diffusion 3 - SD3 Not loading

  #3693 closed Jun 12, 2024

• SD3 Error - RuntimeError: cuDNN Frontend error: s_kv not a multiple of 64 with dropout enabled is not supported with cudnn version below 9.0.0

  #3691 closed Jun 12, 2024

• SRFormer Support?

  #2698 closed Jun 12, 2024

• [Feature Request] Implement Rectified Flow inference for SD3

  #3694 closed Jun 12, 2024

• SD3: 'NoneType' object has no attribute 'tokenize'

  #3690 closed Jun 12, 2024

• Sticky Mouse On Failed Nodes

  #3614 closed Jun 12, 2024

• After reinstallation can't launch Comfy

  #3687 closed Jun 11, 2024

• When running K-sampler: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

  #3635 closed Jun 11, 2024

36 Issues opened by 33 people

• Potential bug in sd3_clip : encode_token_weights (depending how Stability trained it)

  #3772 opened Jun 18, 2024

• I updated ComfyUI and now CheckpointLoaderSimple is nowhere to be found!!!

  #3771 opened Jun 18, 2024

• ksample runtime error

  #3770 opened Jun 18, 2024

• lora+checkpoint producing different results during saving

  #3768 opened Jun 18, 2024

• How to delete the image cache in the image loading?

  #3767 opened Jun 18, 2024

• How to remove custom nodes through the UI manager?

  #3764 opened Jun 17, 2024

• Error window needs a scrollbar

  #3762 opened Jun 17, 2024

• Error window needs a scrollbar

  #3761 opened Jun 17, 2024

• Can no longer copy/paste from one workflow in a browser tab to another workflow in another browser tab

  #3760 opened Jun 17, 2024

• Can't connect small nodes

  #3757 opened Jun 17, 2024

• [ComfyUI-Manager] Starting dependency installation/(de)activation for the extension

  #3756 opened Jun 17, 2024

• no work

  #3755 opened Jun 17, 2024

• Nodes that have failed to load will show as red on the graph.

  #3753 opened Jun 17, 2024

• Request for ComfyUI support for Hunyuan DIT

  #3751 opened Jun 17, 2024

• GUI gets progressively laggy over time.

  #3748 opened Jun 16, 2024

• SD3 - Mac M1 - Picture Quality (Oversaturated/Noisy)

  #3747 opened Jun 16, 2024

• SD3 checkpoint in diffusers format failing to be loaded with Load Checkpoint node

  #3745 opened Jun 16, 2024

• An Idea / Question💡

  #3743 opened Jun 16, 2024

• An error on KSampler

  #3742 opened Jun 16, 2024

• Backend Read-Only/Restricted/Frozen Mode?

  #3740 opened Jun 16, 2024

• Round trip metadata

  #3739 opened Jun 16, 2024

• [Feature]Support for realtime audio input, microphone

  #3737 opened Jun 16, 2024

• Connection errors makeing installing nodes fail + The custom node DB is currently being updated

  #3736 opened Jun 15, 2024

• 'TextField' object has no attribute 'print_text'

  #3735 opened Jun 15, 2024

• [Feature] Support for SD3 ControlNet

  #3734 opened Jun 15, 2024

• Error occurred when executing CLIPTextEncode in sd3_medium_incl_clips_t5xxlfp8

  #3725 opened Jun 14, 2024

• Workflow is stuck for inpainting

  #3722 opened Jun 14, 2024

• The UI choice of SD3 nodes

  #3721 opened Jun 14, 2024

• Access-Control-Allow-Headers does not contain Comfy-User

  #3720 opened Jun 14, 2024

• Ksample not working,how can i do?

  #3714 opened Jun 13, 2024

• loading fbgemm error. fbgemm.dll or one of its dependencies.

  #3703 opened Jun 13, 2024

• ModuleNotFoundError: No module named 'comfy_extras.nodes_model_merging'

  #3702 opened Jun 13, 2024

• Lora keys not loaded for SD3-Medium Lora

  #3701 opened Jun 13, 2024

• feature request - align your steps support SD3

  #3695 opened Jun 12, 2024

• Error first try SD3 directml RX580

  #3689 opened Jun 12, 2024

• After updating the latest version, a BUG will appear in the prompt queue.

  #3686 opened Jun 11, 2024

34 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Allow bfloat16 computations on compatible CPUs with Intel Extension for PyTorch

  #3649 commented on Jun 17, 2024 • 9 new comments

• Execution Model Inversion

  #2666 commented on Jun 17, 2024 • 9 new comments

• After the last update black images after few steps on Mac mini M2 pro with Sonoma 14.5

  #3521 commented on Jun 18, 2024 • 9 new comments

• Add support for the revolutionary B-Lora - Content and Style Separation

  #3674 commented on Jun 16, 2024 • 5 new comments

• help 1080ti to run comfy ui -->got query.dtype: struct c10::Half key.dtype: float and value.dtype: float

  #3593 commented on Jun 15, 2024 • 4 new comments

• Cannot handle this data type: (1, 1, 3), <f4

  #3539 commented on Jun 13, 2024 • 4 new comments

• Allow for nodes to define whether they default to a widget or input

  #3205 commented on Jun 13, 2024 • 2 new comments

• add avif parsing function

  #3565 commented on Jun 18, 2024 • 2 new comments

• Add basic zluda support

  #2829 commented on Jun 18, 2024 • 2 new comments

• [Feature request] Make reroutes more compact

  #977 commented on Jun 18, 2024 • 2 new comments

• [AMD ROCm] No reactions for 3 monthes or more

  #3609 commented on Jun 16, 2024 • 2 new comments

• Support for Lumina T2X

  #3458 commented on Jun 15, 2024 • 2 new comments

• Slot Name will auto reset every time.

  #3580 commented on Jun 11, 2024 • 1 new comment

• Fix to #3593 - attention_pytorch and attention_xformers not properly handling the --dont-upcast-attention flag

  #3637 commented on Jun 16, 2024 • 1 new comment

• feat: add tcd sampler

  #3370 commented on Jun 13, 2024 • 1 new comment

• [Feature Request] Attention Mask for LORAs?

  #2445 commented on Jun 11, 2024 • 1 new comment

• - Required input is missing: ckpt_name | Undefined ckpt_name

  #1172 commented on Jun 11, 2024 • 1 new comment

• Model is still re-loading even though --highvram is turned on

  #3423 commented on Jun 18, 2024 • 1 new comment

• Who are Mat1 and Mat2 and why won't they leave me alone?

  #3102 commented on Jun 18, 2024 • 1 new comment

• Wrong sigma on mac mps backend.

  #3346 commented on Jun 12, 2024 • 1 new comment

• --listen issue for local network

  #3646 commented on Jun 16, 2024 • 1 new comment

• VAE encode/decode messing up faces

  #673 commented on Jun 16, 2024 • 1 new comment

• T shapes cannot be multiplied.

  #3210 commented on Jun 12, 2024 • 1 new comment

• Input types 'tensor<1x257x1xf16>' and 'tensor<1xf32>' are not broadcast compatible

  #3084 commented on Jun 15, 2024 • 1 new comment

• Serious network issues before being capable to get the even first wallpaper done

  #3026 commented on Jun 15, 2024 • 1 new comment

• can't load lcm checkpoint, lcm lora works well

  #1933 commented on Jun 15, 2024 • 1 new comment

• Comfyui shows high-quality previews?

  #3551 commented on Jun 14, 2024 • 1 new comment

• When deploying ComfyUI on a fresh Windows installation using Miniconda, I encountered the "1Torch was not compiled with flash attention" warning during the initial inference.

  #3265 commented on Jun 14, 2024 • 1 new comment

• Error starting the server: Ports 8188 to 8198 are all in use.

  #3450 commented on Jun 14, 2024 • 1 new comment

• RuntimeError: CUDA error: invalid argument

  #3099 commented on Jun 14, 2024 • 1 new comment

• Creating a loop (not submitting workflow) causes an uninterruptible error.

  #2077 commented on Jun 13, 2024 • 1 new comment

• [AMD ROCm] Always crashes somewhy

  #3574 commented on Jun 13, 2024 • 1 new comment

• Speed up the first run.

  #1992 commented on Jun 13, 2024 • 1 new comment

• Feature: Disable Metadata Toggle

  #2782 commented on Jun 15, 2024 • 0 new comments