NerdyRodent
patreon
New Flux Models!
🕑 Added 2024-11-22 00:40:46 +0000 UTCVideo - https://youtu.be/1AlmHTFWjLk
Black Forest Labs have released a bunch of new models for us to play with, so lets get to it! Depth, Canny, Redux and Fill - plus depth & canny LoRA versions too!
Links
The ComfyUI examples page has the links and locations for each of their workflow, and the ones attached to this post.
Fill - InPainting / Outpainting
The best flux inpainting option yet!
Example images:
Redux - Image Variations
My favourite :) Doesn't need a prompt, but with the timestep variation you can increase your prompt strength.
Depth / Canny
Model or LoRA - they both seem good!
Workflows
Loads of workflow examples this time! Usual device optimisations applied, plus a crop/stitch option for inpainting, and a timestep option to help reign in that redux style a little!
Updates
Redux is loads of fun, so I figured why not have a play with a bit of detail daemon, ultimate upscale, KJNodes "advanced" apply, and mixing in the depth LoRA with Redux too?
Seems to work fine to me! :)
Below is a screenshot showing the creation process in the newly attached, "Flux-Redux-DepthLora_USDU_v1.json" workflow file:
I'm still seeing what strange things can be created, but at least now I can do them in high resolution - and with a bit of depth.
Enjoy!
Comments
Lord Banzai
There are 3 workflows in ComfyUI-LTXVideo
Nerdy Rodent
https://comfyanonymous.github.io/ComfyUI_examples/ltxv/ ;)
Lord Banzai
Anyone else using LTX-Video for ComfyUI?
Nerdy Rodent
Version 0.2.0 is very old! I'd start with upgrading to the current version
Peter Rossa
trying to get redux to work, I got something like this, any idea what Im doing wrong ? # ComfyUI Error Report ## Error Details - **Node Type:** StyleModelLoader - **Exception Type:** Exception - **Exception Message:** invalid style model D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\models\style_models\flux1-redux-dev.safetensors ## Stack Trace ``` File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\execution.py", line 317, in execute output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\execution.py", line 192, in get_output_data return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\execution.py", line 169, in _map_node_over_list process_inputs(input_dict, i) File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\execution.py", line 158, in process_inputs results.append(getattr(obj, func)(**inputs)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\nodes.py", line 969, in load_style_model style_model = comfy.sd.load_style_model(style_model_path) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\FreshSDXL\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 391, in load_style_model raise Exception("invalid style model {}".format(ckpt_path)) ``` ## System Information - **ComfyUI Version:** v0.2.0-2-g00a5d08 - **Arguments:** ComfyUI\main.py --windows-standalone-build - **OS:** nt - **Python Version:** 3.11.8 (tags/v3.11.8:db85d51, Feb 6 2024, 22:03:32) [MSC v.1937 64 bit (AMD64)] - **Embedded Python:** true - **PyTorch Version:** 2.2.1+cu121
Nerdy Rodent
Personally I prefer the BFL ones :)
Runebinder
The Flux Fill CrossStitch workflow is great, thanks :)
EruoAuri
Thanks! Do you kno if there's a big quality difference between the union controlnet and the new xlabs depth and canny models?
Clayfacer
That was quick. Thanks
Clayfacer
need this now. :0
