Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add rougel for cross-device evaluation #768

Open
wants to merge 4 commits into
base: llm
Choose a base branch
from

Conversation

qbc2016
Copy link
Collaborator

@qbc2016 qbc2016 commented Apr 10, 2024

To assess the performance of our fine-tuned model, we leverage the Rouge-L metric and conduct experiments with a large number of clients, utilizing the Dolly-15K dataset as our training corpus.
The Dolly-15K dataset encompasses a total of 15,015 data points, distributed across eight distinct tasks. For a more comprehensive evaluation, we allocate the final task exclusively for evaluation purposes, while dedicating the remaining ones to the training phase.
Our experimental setup involves a network of 200 clients, utilizing a Dirichlet distribution for data partitioning to emulate non-IID conditions across the client base.

To do the evaluation, run

python federatescope/eval/eval_for_rougel/eval.py --cfg federatescope/llm/baselime/xxx.yaml

@qbc2016 qbc2016 requested a review from rayrayraykk April 10, 2024 02:39
@rayrayraykk rayrayraykk changed the title Add rougel for dolly Add rougel for cross-device evaluation Apr 10, 2024
@qbc2016 qbc2016 added ready_for_review [**IMPORTANT**] The pull request is ready for review, which will trigger the unit tests. FS-LLM Federated learning in LLM labels Apr 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
FS-LLM Federated learning in LLM ready_for_review [**IMPORTANT**] The pull request is ready for review, which will trigger the unit tests.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant