search

cocktailpeanut / fluxgym

Dead simple FLUX LoRA training UI with LOW VRAM support
683 stars 44 forks source link

Florence-2 never completes #58

Open fase opened 6 days ago

fase commented 6 days ago

I'm running on a remote server with shared=True enabled. I click "Add AI Captions with Florence-2", it goes to "Processing" and it never finishes. No errors on the console, no timeouts, it just never completes. I'm testing with 2 images right now and just crossed 600+ seconds with no results. CPU is at 100% (python) but it never completes. Would be really nice if you could upload matching txt files as I already have these created with ComfyUI + Florence2.

leoleelxh commented 6 days ago

save issue。And somtimes erro： OSError: Can't load the model for 'multimodalart/Florence-2-large-no-flash-attn'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'multimodalart/Florence-2-large-no-flash-attn' is the correct path to a directory containing a file named pytorch_model.bin, tf_model.h5, model.ckpt or flax_model.msgpack. run_captioning

darkon12 commented 6 days ago

You have to install hf-transfer and timm

leoleelxh commented 6 days ago

You have to install hf-transfer and timm

thanks ， finally I run it on wsl ，and work well！

darkon12 commented 5 days ago

Lucky you. On my system it can't find train.sh I guess I have to change relative path to absolute

Naks92 commented 5 days ago

Hi there.. I'm having the same issue. But I think hf-transfer and timm are indeed installed. pytorch_model.bin: 98%|████████████████████████████████████████▏| 1.51G/1.54G [10:57<00:20, 1.54MB/s] Then I got OSError: Can't load the model for 'multimodalart/Florence-2-large-no-flash-attn'. If you were trying to load it from 'https://huggingface.co/models'...

Metairieman55 commented 4 days ago

I'm on win11. I have 21 images and the Florence-2 took forever and when it finished only the first image was tagged. All other blank. ???

fase commented 3 days ago

You have to install hf-transfer and timm

Already installed. I get "Requirement already satisfied" on all dependencies when I pip install. Thanks for the reply though. Sounds like you helped someone else out.

darkon12 commented 2 days ago

You have to install hf-transfer and timm

Already installed. I get "Requirement already satisfied" on all dependencies when I pip install. Thanks for the reply though. Sounds like you helped someone else out.

Passive-aggressive won't help, try this

pip install https://huggingface.co/spaces/Wauplin/gradio_logsview/resolve/main/gradio_logsview-0.0.5-py3-none-any.whl pip install hf-transfer pip install timm