search

nod-ai / SHARK-Studio

SHARK Studio -- Web UI for SHARK+IREE High Performance Machine Learning Distribution
Apache License 2.0
1.42k stars 171 forks source link

Users\Name\.local directory is created (but not used) even if --local_shark_tank is enabled #1458

Open NeedsMoar opened 1 year ago

NeedsMoar commented 1 year ago

Nothing else to it... the directory shouldn't be created if an alternate directory is used, it spams up the User folder. :D

monorimet commented 1 year ago

Thanks for reporting this. We should be handling file storage more intentionally in general.

Ideally the artifacts generated by the .exe / index.py should be kept within the .exe's directory, aside from the _MEI artifacts from the .exe.

This directory is likely initialized in SHARK downloader somewhere, it should be initializing whichever shark_tank directory is specified by the user / pipelined down from the webui arguments (stable_args.py).

NeedsMoar commented 1 year ago

It seems as though the command line argument works, but then its value is pulled from some default in another module a little while later and the directory created, then it goes back to using the command line argument. 


Using C:\Programs\Shark\cache as local shark_tank cache directory.
gradio temporary image cache located at C:\Programs\Shark\shark_tmp/gradio. You may change this by setting the GRADIO_TEMP_DIR environment variable.
No temporary images files to clear.
vulkan devices are available.
metal devices are not available.
cuda devices are not available.
local-sync devices are available.
local-task devices are available.
diffusers\models\cross_attention.py:30: FutureWarning: Importing from cross_attention is deprecated. Please import from diffusers.models.attention_processor instead.
Running on local URL:  http://0.0.0.0:8080

To create a public link, set `share=True` in `launch()`.
shark_tank local cache is located at C:\Users\NicholasCage\.local/shark_tank/ . You may change this by setting the --local_tank_cache= flag
Found device AMD Radeon RX 7900 XTX. Using target triple rdna3-7900-windows.
Using tuned models for CompVis/stable-diffusion-v1-4(fp16) on device vulkan://00000000-2300-0000-0000-000000000000.
NeedsMoar commented 1 year ago

That last line is where it crashed with no message, which is the other bug I just filed. :D

La-Cabra commented 1 year ago

The cause for the --local_tank_cache= flag bug is somewhere inside the code of the tkinter UI. The web interface doesn't have this problem.

The flag for reverting to web interface is: --ui=web.

NeedsMoar commented 1 year ago

Trying to find a working build newer than last release since 787 and 797 both crash with the current AMD drivers (23.7.1), and found this going on in the current build when the web UI is forced; somehow we've got 3 instances of trying to use the wrong directory now and it's back to being created...

Using C:\Programs\Shark\cache as local shark_tank cache directory. gradio temporary image cache located at C:\Programs\Shark\shark_tmp/gradio. You may change this by setting the GRADIO_TEMP_DIR environment variable. No temporary images files to clear. vulkan devices are available. shark_tank local cache is located at C:\Users\Ѽ.local/shark_tank/ . You may change this by setting the --local_tank_cache= flag metal devices are not available. cuda devices are not available. local-sync devices are available. shark_tank local cache is located at C:\Users\Ѽ.local/shark_tank/ . You may change this by setting the --local_tank_cache= flag local-task devices are available. shark_tank local cache is located at C:\Users\Ѽ.local/shark_tank/ . You may change this by setting the --local_tank_cache= flag