Add tutorial for mlc-llm model compilation in python #13
Conversation
Co-authored-by: CharlieRuan0130 <[email protected]> Co-authored-by: sbelcmu <[email protected]>
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
| "name": "stdout", | ||
| "text": [ | ||
| "Using path \"Llama-2-7b-chat-hf\" for model \"Llama-2-7b-chat-hf\"\n", | ||
| "WARNING: --db-path does not point to a valid database: []\n", |
There was a problem hiding this comment.
I checked that the default value of --db-path is still "log_db" in BuildArgs https://github.com/mlc-ai/mlc-llm/blob/d4c3a17379e1e890db6650568d7b5a59650a17bd/mlc_llm/core.py#L60. So just a bit curious why it emits this warning.
There was a problem hiding this comment.
Ohh I think this is because we are doing the model building in my_workspace instead of inside the mlc_llm repo. The intent is to show that, by installing the mlc_llm as a package, they can build wherever they want.
There was a problem hiding this comment.
I see. Sounds good for now, as we will remove the database soon when dlight gets maturer.
There was a problem hiding this comment.
Makes sense. Thanks for the suggestions!
There was a problem hiding this comment.
Looks very pretty! Thanks @CharlieRuan0130 and @sbelcmu for the notebook. Left two comments above and we are good to go after addressing them.
Co-authored-by: Ruihang Lai <[email protected]>
No description provided.