Some errors and solutions (May, 2025)

[Taylor-Swift1]

cuda version: 12.8 system: Ubuntu 20.04.6 LTS

• Problem: when I try to build flash-attention by "pip install . --no-build-isolation". note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for flash-attn Running setup.py clean for flash-attn Failed to build flash-attn ERROR: Failed to build installable wheels for some pyproject.toml based projects (flash-attn) The latest version of flash-attention (now) is v2.7.4.post1, which drops Pytorch 2.1. This is conflicted with the required version of pytorch. Solution: run "pip install flash-attention==2.7.3 --no-build-isolation" (no 2.7.4 for some reasons).

• Problem: When I prepare InstructDiffusion checkpoint, get an error to convert files ImportError: cannot import name 'cached_download' from 'huggingface_hub' The lastet huggingface_hub package doesn't support this function. Solution: try to install a low version of huggingface_hub. Here I solved by "pip install huggingface_hub==0.24.0"

If I have more problems and solved them, will post them here.

[Taylor-Swift1]

• Problem: no LLaVa v1.1 Solution: https://huggingface.co/liuhaotian/LLaVA-13b-delta-v1-1

[fwbc]

• pip install flash-attention==2.7.3 --no-build-isolation

run pip install flash-attention==2.7.3 --no-build-isolation but "Could not find a version that satisfies the requirement flash-attention==2.7.3 (from versions: 1.0.0)"

[Taylor-Swift1]

@fwbc I would suggest to check the versions that you have. Then you can choose the highest.

[Taylor-Swift1]

Traceback (most recent call last): File "/data2/zg5/image_edit/smart_edit/SmartEdit/test/DS_SmartEdit_test.py", line 169, in <module> from model.DS_SmartEdit_model import SmartEdit_model ModuleNotFoundError: No module named 'model'

Add this to test file import sys import os sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))

[Taylor-Swift1]

no module basicsr

pip install basicsr

[fwbc]

cuda version: 12.8 system: Ubuntu 20.04.6 LTS

• Problem: when I try to build flash-attention by "pip install . --no-build-isolation". note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for flash-attn Running setup.py clean for flash-attn Failed to build flash-attn ERROR: Failed to build installable wheels for some pyproject.toml based projects (flash-attn) The latest version of flash-attention (now) is v2.7.4.post1, which drops Pytorch 2.1. This is conflicted with the required version of pytorch. Solution: run "pip install flash-attention==2.7.3 --no-build-isolation" (no 2.7.4 for some reasons).
• Problem: When I prepare InstructDiffusion checkpoint, get an error to convert files ImportError: cannot import name 'cached_download' from 'huggingface_hub' The lastet huggingface_hub package doesn't support this function. Solution: try to install a low version of huggingface_hub. Here I solved by "pip install huggingface_hub==0.24.0"

If I have more problems and solved them, will post them here.

python test/DS_SmartEdit_test.py --is_understanding_scenes True --model_name_or_path "./checkpoints/vicuna-7b-v1-1" --LLaVA_model_path "./checkpoints/LLaVA-7B-v1" --save_dir './checkpoints/SmartEdit-7B/Understand-15000' --steps 15000 --total_dir "./checkpoints/SmartEdit-7B" --sd_qformer_version "v1.1-7b" --resize_resolution 256 python test/DS_SmartEdit_test.py --is_reasoning_scenes True --model_name_or_path "./checkpoints/vicuna-7b-v1-1" --LLaVA_model_path "./checkpoints/LLaVA-7B-v1" --save_dir './checkpoints/SmartEdit-7B/Reason-15000' --steps 15000 --total_dir "./checkpoints/SmartEdit-7B" --sd_qformer_version "v1.1-7b" --resize_resolution 256 python test/DS_SmartEdit_test.py --is_understanding_scenes True --model_name_or_path "./checkpoints/vicuna-13b-v1-1" --LLaVA_model_path "./checkpoints/LLaVA-13B-v1" --save_dir './checkpoints/SmartEdit-13B/Understand-15000' --steps 15000 --total_dir "./checkpoints/SmartEdit-13B" --sd_qformer_version "v1.1-13b" --resize_resolution 256 python test/DS_SmartEdit_test.py --is_reasoning_scenes True --model_name_or_path "./checkpoints/vicuna-13b-v1-1" --LLaVA_model_path "./checkpoints/LLaVA-13B-v1" --save_dir './checkpoints/SmartEdit-13B/Reason-15000' --steps 15000 --total_dir "./checkpoints/SmartEdit-13B" --sd_qformer_version "v1.1-13b" --resize_resolution 256 do you know where are these models?I did not find like "vicuna-7b-v1-1" on their hugging face.And because their huggingface did not provide additional user guides and notebooks,I do not know how to use their model.

[baihuple]

you can use the release version. Just do the command "pip install xxx.whl" and hope it'll help you!