requirements.txt with deps fixed to specific versions

[smartinezbragado]

Hello,

I'm attempting to deploy Flux LoRA training on SageMaker but I'm encountering numerous dependency issues that are difficult to debug. Does anyone have a requirements.txt file with the dependencies fixed to a specific version?

Thank you in advance!

[Clement-Lelievre]

Hello,

I'm attempting to deploy Flux LoRA training on SageMaker but I'm encountering numerous dependency issues that are difficult to debug. Does anyone have a requirements.txt file with the dependencies fixed to a specific version?

Thank you in advance!

hi,

here the one I use:

accelerate==0.33.0
albumentations==1.4.3
bitsandbytes==0.43.3
controlnet_aux==0.0.7
git+https://github.com/huggingface/diffusers.git@6dc6486565ea1d8d1be567eefc1094e9185560a1
einops==0.8.0
flatten_json==0.1.14
hf_transfer==0.1.8
huggingface-hub==0.24.5
k-diffusion==0.1.1.post1 
kornia==0.7.3
lpips==0.1.4
lycoris-lora==1.8.3
omegaconf==2.3.0
open_clip_torch==2.26.1
opencv-python==4.9.0.80
optimum-quanto==0.2.4
oyaml==1.0
pillow==10.2.0
prodigyopt==1.0
pydantic==1.10.17
pyyaml==6.0.1
requests==2.32.3
safetensors==0.4.4
scikit-image==0.24.0
sentencepiece==0.2.0
tensorboard==2.17.0
timm==1.0.8
toml==0.10.2
torch==2.4.1
torchvision==0.19.1
transformers==4.44.0

[inflamously]

Interesting to note is that I had only optimum-quanto==0.2.4 work under windows with torch 2.4.1+cu124 and CUDA 12.4 installed for me. When trying to use newly released optimum-quanto==0.2.5, ninja starts to throw various exceptions about not finding __asm while compiling.

Edit: Windows 11 Pro 24H2 used

[HelixNGC7293]

Interesting to note is that I had only optimum-quanto==0.2.4 work under windows with torch 2.4.1+cu124 and CUDA 12.4 installed for me. When trying to use newly released optimum-quanto==0.2.5, ninja starts to throw various exceptions about not finding __asm while compiling.

Edit: Windows 11 Pro 24H2 used

I got the exactly same issue, after change to optimum-quanto==0.2.4 everything works fine! Thanks inflamously!