2462 Commits (f7a5107784cded39f92a4bb7553507575e78edbe)
 

Author SHA1 Message Date
comfyanonymous a3dffc447a Support AuraFlow Lora and loading model weights in diffusers format.
You can load model weights in diffusers format using the UNETLoader node.
7 months ago
comfyanonymous ce2473bb01 Add link to AuraFlow example in Readme. 8 months ago
Robin Huang 4ca9b9cc29
Add Github Workflow for releasing stable versions and standalone bundle. (#3949)
* Add stable release.

* Only build CUDA 12.1 + 3.11 Python.

* Upgrade checkout and setup-python to latest version.

* lzma2

* Update artifact name to be ComfyUI_windows_portable_nvidia.7z
8 months ago
comfyanonymous 29c2e26724 Better tokenizing code for AuraFlow. 8 months ago
comfyanonymous b6f09cf649 Add sentencepiece dependency. 8 months ago
comfyanonymous 8e012043a9 Add a ModelSamplingAuraFlow node to change the shift value.
Set the default AuraFlow shift value to 1.73 (sqrt(3)).
8 months ago
comfyanonymous 9f291d75b3 AuraFlow model implementation. 8 months ago
comfyanonymous f45157e3ac Fix error message never being shown. 8 months ago
comfyanonymous 5e1fced639 Cleaner support for loading different diffusion model types. 8 months ago
comfyanonymous ffe0bb0a33 Remove useless code. 8 months ago
comfyanonymous 391c1046cf More flexibility with text encoder return values.
Text encoders can now return other values to the CONDITIONING than the cond
and pooled output.
8 months ago
comfyanonymous e44fa5667f Support returning text encoder attention masks. 8 months ago
Chenlei Hu 90389b3b8a
Update bug issue template (#3996)
* Update issue template

* nit
8 months ago
Chenlei Hu 8d3f979b63
Check unhandled exception in test log in test action (#3987)
* Upload console logs

* Check unhandled exception
8 months ago
Chenlei Hu 83f70a88fb
Add __module__ to node info (#3936)
Use more explicit name 'python_module'

Parse abs ath

Move parse to nodes.py
8 months ago
Extraltodeus f1a01c2c7e
Add sampler_pre_cfg_function (#3979)
* Update samplers.py

* Update model_patcher.py
8 months ago
comfyanonymous c3db344746 Fix ConditioningZeroOut when there is no pooled output. 8 months ago
bymyself d160073829
Fix loadGraphData call during restore (#3976) 8 months ago
comfyanonymous ade7aa1b0c Remove useless import. 8 months ago
comfyanonymous faa57430b0 Controlnet union model basic implementation.
This is only the model code itself, it currently defaults to an empty
embedding [0] * 6 which seems to work better than treating it like a
regular controlnet.

TODO: Add nodes to select the image type.
8 months ago
comfyanonymous bb663bcd6c Rename clip_t5base to t5base for stable audio text encoder. 8 months ago
comfyanonymous 628f0b8ebc Move audio nodes out of _for_testing. 8 months ago
comfyanonymous 2dc84d1444 Add a way to set the timestep multiplier in the flow sampling. 8 months ago
comfyanonymous ff63893d10 Support other types of T5 models. 8 months ago
comfyanonymous 4040491149 Better T5xxl detection. 8 months ago
comfyanonymous b8e58a9394 Cleanup T5 code a bit. 8 months ago
comfyanonymous 80c4590998 Allow specifying the padding token for the tokenizer. 8 months ago
comfyanonymous ce649d61c0 Allow zeroing out of embeds with unused attention mask. 8 months ago
comfyanonymous b4c2d03d47 Remove duplicate import. 8 months ago
comfyanonymous 1dc87df4c5 Readme changes. 8 months ago
comfyanonymous cedbc94cc0 Forgot this in last commit. 8 months ago
comfyanonymous bd2d3e27d7 Show comfy_extras warning at the end.
Remove code.
8 months ago
comfyanonymous 720b17442d Temporary revert. 8 months ago
Chenlei Hu 0e3dfd9e34
Use relative path for custom/extra node module name (#3944)
* Fix module name for comfy extra nodes

* Use module name relative to root dir
8 months ago
comfyanonymous 739b76630e Remove useless code. 8 months ago
bymyself 24b969d3da
Skip state check hook on first load (#3915) 8 months ago
Chenlei Hu 086ac75228
3.8 Compatible type annotation (#3938) 8 months ago
comfyanonymous d7484ef30c Support loading checkpoints with the UNETLoader node. 8 months ago
comfyanonymous 537f35c7bc Don't update dict if contiguous. 8 months ago
Alex "mcmonkey" Goodwin 3f46362d22
fix non-contiguous tensor saving (from channels-last) (#3932) 8 months ago
comfyanonymous 01991f72ce Fix SamplerEulerCFGpp node. 8 months ago
comfyanonymous 2f03201690 Remove some empty lines. 8 months ago
shawnington 52aaee251f
Fix to #3465. Prevent, resaving of duplicate images if overwrite not specified (#3472)
* Fix to #3465. Prevent the, resaving of duplicate images if overwrite not specified

This is a fix to #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
8 months ago
Bob Du 1ef66b0955
Add example for how to add custom API routes (#3597) 8 months ago
Chenlei Hu 9dd549e253
Add `--no-custom-node` cmd flag (#3903)
* Add --no-custom-node cmd flag

* nit
8 months ago
Peter Crabtree b82d67d5bf
Add SamplerEulerAncestralCFG++ custom sampler node (#3901)
(for eta and s_noise)
8 months ago
Hayden Reeve 755c48d78e
Fix several typos in example_node.py.example (#3204)
This change includes corrections for several spelling errors in the
documentation of example_node.py.example file.

These were previously raised by #3157, but they missed a few.
8 months ago
comfyanonymous 5dccfefe8d Switch nightly pytorch standalone package to lzma2. 8 months ago
YAN Wenkun 0cd4a6a5e5
Fine-tuning GitHub Actions (#3169)
* Bumping GitHub Actions versions

* Using LZMA2 for 7zip compression in Windows packaging
8 months ago
Robin Huang 601b4b63e1
Add CONTRIBUTING.md (#3910)
* Create CONTRIBUTING.md

* Add feature-request channel link.

* Remove discord links for channels.
8 months ago