Scripts for fine-tuning Meta Llama with composable FSDP & PEFT methods to cover single/multi-node GPUs. Supports default & custom datasets for applications such as summarization and Q&A. Supporting a number of candid inference solutions such as HF TGI, VLLM for local or cloud deployment. Demo apps to showcase Meta Llama for WhatsApp & Messenger.
15.35k
stars
2.22k
forks
source link
Update wandb.py to accept setting run name from command line argument (e.g., --wandb_config.name "run_name") for fine tuning #772
Current wandb logging generate a random run name (despite wandb supporting assigning a run name); I added a missing name attribute to wandb_config, so one could use command line argument like this --wandb_config.name "run_name" to control the displayed run_name on the wandb webpage.
Feature/Issue validation/testing
Before and after the change of this commit, and use argument --wandb_config.name "fake run name". Note that stoic-eon-2 is a random name.
Before submitting
[ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
What does this PR do?
Current wandb logging generate a random run name (despite wandb supporting assigning a run name); I added a missing name attribute to wandb_config, so one could use command line argument like this
--wandb_config.name "run_name"to control the displayed run_name on the wandb webpage.Feature/Issue validation/testing
Before and after the change of this commit, and use argument --wandb_config.name "fake run name". Note that
stoic-eon-2is a random name.Before submitting