Jesse White
Jesse White
This has been open without changes for 6 weeks. Can we merge this in @joecummings?
Can you update the readme with an example for MPS?
checking in again to see what the plan is on merging. Still haven't seen any movement. Would love to see mps make it to main.
what steps did you take to get this message? I'm not seeing it on my end.
I can test it on my machine but i need details on which commands to run (general instructions). We need this in general for this readme. Either the main readme.md...
I'm unable to re-pro with up to date main
We are currently working on llava support. Getting close! @Gasoonjia @larryliu0820
I haven't tried pure linux since finding this issue on wsl.
Compile also throws a similar error ``` python3 torchchat.py generate llama3.1 --prompt "tell me a story about a bear Using device=cuda NVIDIA GeForce RTX 4090 Time to load model: 8.90...
> This looks like a WSL lib path issue. The compilation cmd is looking for `Python.h` in `-I/usr/include/python3.11`, which is probably wrong. cc @xuhancn who has more experience on Windows...