Updated Conda packages and channels to install cuda-toolkit and override 12.0 cuda packages requested by pytorch with their 11.7 equivalent.
Removed Conda installation since we can use the downloaded Micromamba.exe for the same purpose with a smaller footprint.
Removed redundant PATH changes.
Changed %gpuchoice% comparisons to be case-insensitive.
Added additional error handling and removed the use of .tmp files.
Added missing extension requirements.
Added GPTQ installation. Will attempt to compile locally and, if failed, will download and install a precompiled wheel.
Incorporated fixes from one-click-bandaid.
Fixed and expanded first sed command from one-click-bandaid.
libbitsandbytes_cudaall.dll is used here as the cuda116.dll used by one-click-bandaid does not work on my 1080ti. This can be changed if needed.
Now shows the message in the console when unloading weights. Also reload_model() calls unload_model() first to free the memory so that multiple reloads won't overfill it.
Reworked remove_surrounded_chars() to use regular expression ( https://regexr.com/7alb5 ) instead of repeated string concatenations for elevenlab_tts, silero_tts, sd_api_pictures. This should be both faster and more robust in handling asterisks.
Reduced the memory footprint of send_pictures and sd_api_pictures by scaling the images in the chat to 300 pixels max-side wise. (The user already has the original in case of the sent picture and there's an option to save the SD generation).
This should fix history growing annoyingly large with multiple pictures present
There doesn't seem to be much value to them - they just take up space while also making it seem like there's still some sort of pseudo-dialogue going on, instead of a monologue by the bot.
