-
Notifications
You must be signed in to change notification settings - Fork 52
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
Smurty version #170
base: main
Are you sure you want to change the base?
Smurty version #170
Conversation
|
||
|
||
def get_action_post_hoc(agent: GenericAgent, step_info): | ||
""" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think this can be safely ported to the unsupervised project
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we can remove this too. This is some old stuff i was trying
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am not sure about what these changes do. Worst case you could copy this file and run it from your laptop. We could also make changes to the main x-ray if they are generic use cases
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There were some issues with the earlier version of agentlab, which this fixes. But now, i think we can discard these
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If this is specific to your use case, we could bring this to the unsupervised project and change the import in the so that the modified file is imported
@@ -20,6 +20,7 @@ def run_experiments( | |||
exp_args_list: list[ExpArgs], | |||
study_dir, | |||
parallel_backend="joblib", | |||
save_transition_history=False, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
saving transition history should be done by placing them inside the AgentInfo. I -and @ThibaultLSDC - can help you figure this out :)
except FileNotFoundError: | ||
yield exp_result.exp_args | ||
# yield exp_result.exp_args | ||
try: | ||
yield exp_result.exp_args | ||
except Exception as e: | ||
logging.error(f"Error with exp_result.exp_args: {e}") |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is not needed
def run( | ||
self, | ||
n_jobs=1, | ||
parallel_backend="joblib", | ||
strict_reproducibility=False, | ||
): | ||
"""Run all experiments in the study in parallel when possible. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this is just cosmetic
run_experiments( | ||
n_jobs, | ||
self.exp_args_list, | ||
self.dir, | ||
parallel_backend=parallel_backend, | ||
) | ||
report_df = self.get_report(ignore_cache=True) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
just cosmetic
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@ThibaultLSDC @recursix Would VLLMChatModels be relevant for AgentLab?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This file has been deprecated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thibault made a PR to port this to AgentLab main
No description provided.