Open C0nsumption opened 1 week ago
Which python version did you used. In the conda environment it is explicitly said 3.9
But did you use the last version of python for the virtual env ?
Which python version did you used. In the conda environment it is explicitly said 3.9
But did you use the last version of python for the virtual env ?
3.10.9
My bad for the delay bud, no github notification. Think it only comes up when quoted. But honestly if you are just doing inference my assumption is you should be fine even with 3.11 and maybe 3.12. Cause all the errors that pop up without these changes are all training related. It has nothing to do with actual inference.
Windows?
feat: Add Compatibility for Windows
Description
This pull request introduces several changes to ensure compatibility with Windows and the most recent versions of various modules. The following modifications have been made:
Module Upgrades:
torch
,diffusers
, etc.Downgrade NumPy:
Conditional Import and Usage of
deepspeed
:deepspeed
and add its related arguments only ifdeepspeed
is available.ImportError
and makes the script more robust.Enhance Image-Saving Logic:
get_next_index
function to handle non-integer filenames during the image-saving process.Installation Instructions
Clone the Repository:
Install Dependencies:
Using a Virtual Environment
Note: These versions are not the ones used in the
requirements.txt
, but they allow the use of CUDA 12.1 and the newest versions ofdiffusers
andpytorch
. (Same as ComfyUI I believe)Using Conda for Closer Compatibility with Original Repository
Run Inference:
WHY?
Error Encountered
flash_attn
: No module named 'flash_attn'deepspeed
: No module named 'deepspeed'Solution
Downgrade NumPy:
Changes in
hydit/modules/models.py
:args.use_flash_attn
. If the attribute does not exist, it returnsFalse
instead of raising anAttributeError
. This maintains the intended behavior without causing an error.Changes in
hydit/config.py
:deepspeed
and add its related arguments only ifdeepspeed
is available. This ensures thatdeepspeed
related arguments are only included whendeepspeed
is successfully imported, avoiding potentialImportError
and making the script more robust.Changes in
sample_t2i.py
:get_next_index
function to handle non-integer filenames gracefully during the image-saving process. This ensures that the script can handle filenames that do not conform to an integer pattern without breaking.Testing
Known Bugs
deepspeed
andflash_attn
are currently not supported and training is not supported.