I upgraded from 0.2.31 to 0.3.2 and there are several issues that make it feel like a downgrade.
The garbage can icon for deleting chats should appear when you hover over them like before and there should be a prompt. Having to click the three dots and then delete every time is unintuitive and caused me to delete my whole Migrated chats folder when I intended to delete the latest chat above it.
I recovered the files with an UNELETE app but they're either corrupted or I'm doing something wrong because they all appear as Unknown chats when I fire up LM studio. That's the jsons in the conversation folder. In the chats folder, the contents seem preserved when I open them in text editor but copying these to the conversation folder also did nothing.
How do I restore these to how they were before my mistake?
It no longer shows the token count when I type or copy-paste stuff into the chat. You guys need to bring this back, it's crucial for me to know how close to the context window I am when I copy-paste an entire novel/post/document. Sure, it shows the token count AFTER each prompt but I need to know how much the initial paste will be.
Why does it only allow me to select half the threads my hardware is capable of using? I have a 6-core CPU with hyperthreading and the settings now only allow me to select a maximum of 3 when I could select 6 before. The performance didn't take a huge hit but I notice increased lag between the responses.
Why does the app continue processing after it finishes a response to my prompt? I try to press send after typing my response but it says I have to wait for it to finish despite that it did finish. I can press "stop generating" and do it but this seems to cause a greater delay when the LLM responds next. The previous version of LM Studio I used did not do this.
Why can we no longer save our settings as a profile? It's annoying having to update the system prompt for every single new chat I start.
I upgraded from 0.2.31 to 0.3.2 and there are several issues that make it feel like a downgrade.
I recovered the files with an UNELETE app but they're either corrupted or I'm doing something wrong because they all appear as Unknown chats when I fire up LM studio. That's the jsons in the conversation folder. In the chats folder, the contents seem preserved when I open them in text editor but copying these to the conversation folder also did nothing. How do I restore these to how they were before my mistake?
It no longer shows the token count when I type or copy-paste stuff into the chat. You guys need to bring this back, it's crucial for me to know how close to the context window I am when I copy-paste an entire novel/post/document. Sure, it shows the token count AFTER each prompt but I need to know how much the initial paste will be.
Why does it only allow me to select half the threads my hardware is capable of using? I have a 6-core CPU with hyperthreading and the settings now only allow me to select a maximum of 3 when I could select 6 before. The performance didn't take a huge hit but I notice increased lag between the responses.
Why does the app continue processing after it finishes a response to my prompt? I try to press send after typing my response but it says I have to wait for it to finish despite that it did finish. I can press "stop generating" and do it but this seems to cause a greater delay when the LLM responds next. The previous version of LM Studio I used did not do this.
Why can we no longer save our settings as a profile? It's annoying having to update the system prompt for every single new chat I start.