How can I go back to previous version

[lomeiyemei]

I got error while auto update, so I freshly install v2.5.0. It can run but However I cannot have my picture outpaint successfully done so far. What is the method to restore back to previous version?

[mashb1t]

What's the issue with outpaint?

You can not easily jump back, as you also have to downgrade all packages again. What you can do is

• download the zip file of the first release again, unzip it using 7zip, manually install git, run git checkout v2.4.3 on a terminal and modify your run.bat to use launch.py instead of entry_with_update.py, then execute run.bat.
• OR use the docker image 2.4.3.

[lomeiyemei]

I got this error from terminal. I am running on a low-end system.

Traceback (most recent call last):
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\async_worker.py", line 1181, in handler
denoising_strength, initial_latent, width, height, current_progress = apply_inpaint(async_task,
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\async_worker.py", line 484, in apply_inpaint
inpaint_worker.current_task = inpaint_worker.InpaintWorker(
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\inpaint_worker.py", line 162, in init
self.interested_image = perform_upscale(self.interested_image)
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\upscaler.py", line 30, in perform_upscale
img = opImageUpscaleWithModel.upscale(model, img)[0]
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\ldm_patched\contrib\external_upscale_model.py", line 54, in upscale
s = ldm_patched.modules.utils.tiled_scale(in_img, lambda a: upscale_model(a), tile_x=tile, tile_y=tile, overlap=overlap, upscale_amount=upscale_model.scale, pbar=pbar)
File "F:\fooocus\Fooocus_win64_2-5-0\python_embeded\lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\ldm_patched\modules\utils.py", line 418, in tiled_scale
ps = function(s_in).to(output_device)
RuntimeError: CUDA error: the launch timed out and was terminated
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\async_worker.py", line 1440, in worker
handler(task)
File "F:\fooocus\Fooocus_win64_2-5-0\python_embeded\lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "F:\fooocus\Fooocus_win64_2-5-0\python_embeded\lib\site-packages\torch\utils_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "F:\fooocus\Fooocus_win64_2-5-0\Fooocus\modules\async_worker.py", line 1193, in handler
except EarlyReturnException:
TypeError: catching classes that do not inherit from BaseException is not allowed
Total time: 34.91 seconds

[mashb1t]

that's actually a bug in Fooocus. Can be prevented by disabling debugging of the inpaint preprocessor for now, will fix tomorrow.

[lomeiyemei]

thanks for your prompt reply and appreciate for your work

[mashb1t]

@lomeiyemei fixed in main (PR #3332), please restart Fooocus again so it can pull the update