32 commits

Author	SHA1	Message	Date
Isadora White	2f80b65d42	fixing small bugs related to single agent support	2025-03-09 13:01:54 -07:00
Isadora White	3576c6fb07	Merge branch 'main' into main_save_03-07-2025	2025-03-09 12:02:58 -07:00
Isadora White	e9e8cf3c88	single agent task support in evaluation script	2025-03-09 12:02:48 -07:00
Isadora White	b75d941d97	smol changes	2025-03-08 19:24:13 -08:00
Isadora White	7492582ed9	changed eval script slightly	2025-03-07 17:29:43 -08:00
Mehul Maheshwari	dda1ba200f	better readability for incoming devs. revised blueprint generation for ladder fix. cheats enabled for jill_0 and andy_0.	2025-03-06 16:16:02 -08:00
Isadora White	9aefb4676a	updating port number for vllm	2025-03-05 15:53:43 -08:00
Isadora White	7e7f893cf3	fix evaluation script bug	2025-03-05 10:10:24 -08:00
Isadora White	6e0d7e1eaf	change evaluation_script to allow for new world names	2025-03-04 22:08:04 -08:00
Isadora White	34145168dc	refactoring changes	2025-03-04 12:09:23 -08:00
Isadora White	e3ec9d34b4	fixing merge related small bugs	2025-03-04 11:54:09 -08:00
mmaheshwari2	2036288e13	tested eval script, changed model to 4o-mini	2025-03-04 10:28:11 -08:00
Isadora White	44fc1b4618	evaluation script for vllm	2025-03-03 06:10:52 +00:00
Isadora White	37b1fc0bed	increase timeout length for adding bot to world for the first time	2025-03-01 21:37:20 -08:00
Isadora White	af79c78fbb	fixing evaluation script to actually add bots as op and add new models	2025-03-01 19:21:35 -08:00
Isadora White	ae39028d3b	longer sleeps, early breaking for scenarios where there is only one agent	2025-02-28 18:31:19 -08:00
Isadora White	39cec7cf82	changed the checking if complete cycle to be more frequent and updated the collab_profile	2025-02-28 16:59:56 -08:00
Isadora White	7cafc254d1	making the default to load in the collaborative profiles	2025-02-23 21:11:08 -08:00
Isadora White	2da97b5607	adding a mechanism to add environment variables to the keys.json automatically	2025-02-23 18:55:13 -08:00
Isadora White	8a75d8a78e	changing the give to player command to account for an edge case where the players are too close together and moving away takes time	2025-02-22 17:53:42 -08:00
Isadora White	719b72da9e	set up to use s3 logging instead of wandb	2025-02-21 17:02:21 -08:00
Isadora White	d4565aa68c	small fixes, the items were being given twice to the agents on initialization and accounting for blocked_actions not being in the task file	2025-02-20 21:45:29 -08:00
Isadora White	7a19f34e22	Merge branch 'main' into evaluation_parallelization	2025-02-18 18:34:46 -08:00
Isadora White	aad19d616c	fixed evaluation script to allow for parallel worlds again	2025-02-18 16:39:31 -08:00
Isadora White	fb5d95debe	fixed the issue with garbling commands by instead putting the commands in a bash script and running them that way	2025-02-17 17:25:12 -08:00
Isadora White	bc15700196	fixed wandb logging	2025-02-15 12:02:44 -08:00
Isadora White	0b2624db6d	updating evaluation script with wandb logs and fixing error in tasks.js	2025-02-13 10:46:44 -08:00
Isadora White	6e48f06e1d	small changes	2025-02-12 15:39:57 -08:00
Ayush Maniar	acf4eece60	Filename includes timestamp of first experiment run	2025-02-11 12:29:29 -08:00
Isadora White	caa5c389e7	logging messages	2025-01-31 12:57:38 -08:00
Isadora White	231204d0c8	using environment variables to help differentiate the settings.js files :D	2025-01-26 11:39:21 -08:00
Ayush Maniar	3bbed21526	Added python script for task evaluation which stores experiment run times and results.	2025-01-23 12:33:21 -08:00