48 Commits

Author	SHA1	Message	Date
hiyouga	5111cac6f8	support report custom args	2024-12-21 21:42:45 +00:00
ZeYi Lin	3a7ea2048a	fix: by hiyouga suggestion	2024-12-20 16:43:03 +08:00
hiyouga	2811814fc4	fix mrope	2024-12-12 15:08:17 +00:00
hiyouga	c38aa29336	support rank0 logger	2024-11-02 18:31:04 +08:00
hiyouga	30567a1487	fix incorrect loss value for vlms	2024-10-30 08:56:46 +00:00
hiyouga	21db8ed2f4	use pre-commit	2024-10-29 09:07:46 +00:00
hiyouga	c7e51ff187	fix #5411	2024-09-11 17:36:42 +08:00
hiyouga	54c6905937	add docstrings, refactor logger	2024-09-08 00:56:56 +08:00
hiyouga	dabad5570b	update get template	2024-09-04 22:36:20 +08:00
hiyouga	47ea97fb1b	lazy image load	2024-09-04 02:27:08 +08:00
hiyouga	9967ccb3ae	fix mixed mm inputs and rlhf-v	2024-09-01 20:52:47 +08:00
hiyouga	3382317e32	refactor mm training	2024-08-30 02:14:31 +08:00
hiyouga	c8b4c7fee5	tiny fix	2024-08-20 00:10:52 +08:00
liu-zichen	ddee718b31	fix lr not change	2024-08-13 16:33:34 +08:00
moontidef	40908a36fa	fix: rename optimzer to optimizer	2024-08-07 10:05:01 +08:00
hiyouga	608de799a2	add unittest	2024-07-19 01:06:27 +08:00
hiyouga	29ebcd75d5	fix up	2024-07-15 01:04:56 +08:00
hoshi-hiyouga	15b399a82f	Merge pull request #4691 from codemayq/feature-suppot-eval-dataset ... add eval dataset support	2024-07-15 01:00:34 +08:00
hiyouga	fb0c400116	fix ppo trainer	2024-07-10 11:05:45 +08:00
codingma	76f3bbcfc0	1. add custom eval dataset support ... 2. merge load dataset and split dataset function	2024-07-05 15:52:10 +08:00
hiyouga	8845e94f91	fix #4609 ... unwrap_model_for_generation(reward_model) is necessary for zero3 training	2024-07-03 19:45:51 +08:00
hiyouga	8b1172b910	tiny fix	2024-07-03 02:31:50 +08:00
hiyouga	71cdf8956e	tiny fix	2024-07-02 23:06:13 +08:00
hiyouga	821bb6660e	remove rlhf support for chatglm2&3	2024-07-02 23:03:17 +08:00
hiyouga	c13ae2df19	upcast logits	2024-07-02 22:32:05 +08:00
hiyouga	c47ab6c072	improve rlhf	2024-07-02 22:23:08 +08:00
hiyouga	4c296001c4	fix ppo callbacks	2024-07-02 17:34:56 +08:00
hiyouga	8baf3b22b0	refactor pissa, improve llamaboard	2024-06-28 01:04:24 +08:00
hiyouga	8d6cd69ac4	fix #4458	2024-06-26 19:52:35 +08:00
hiyouga	095fab58d3	tiny fix about badam	2024-06-25 01:54:53 +08:00
hoshi-hiyouga	d0f953bf5b	Merge pull request #4352 from Ledzy/main ... [Enhancement] Support ZeRO-3 when using BAdam	2024-06-25 01:49:13 +08:00
hiyouga	e507e60638	update readme	2024-06-24 18:22:12 +08:00
Jonery	5c2ff1b749	Cleaner integration.	2024-06-19 12:29:40 +08:00
Jonery	0f72aac8c9	Support distributed BAdam.	2024-06-18 12:27:47 +08:00
hiyouga	38b6b0f52e	tiny fix	2024-06-16 01:06:41 +08:00
hiyouga	d87108daa6	add license	2024-06-15 17:54:33 +08:00
hiyouga	b27269bd2b	add test cases	2024-06-15 04:05:54 +08:00
hiyouga	cf9f2d6c42	fix #4209 ... DeepSpeed ZeRO3 has inflight param error when calling model.eval()	2024-06-13 02:25:50 +08:00
hiyouga	f8d8690bf4	tiny fix	2024-06-07 05:19:21 +08:00
hiyouga	4489d73ac7	fix ppo trainer save zero3 model ... accelerator.get_state_dict(ds_model) should be called at all ranks	2024-06-07 05:14:19 +08:00
hiyouga	2702d7e952	fix ppo in trl 0.8.6	2024-06-07 04:48:29 +08:00
hiyouga	74f96efef9	rename files	2024-06-07 00:09:06 +08:00
hiyouga	76c61905b2	fix ppo+zero3 #3108	2024-06-06 23:30:07 +08:00
hiyouga	149610c636	fix ppo dataset bug #4012	2024-06-06 19:03:20 +08:00
hiyouga	fad2591e31	update trainers	2024-06-06 18:45:49 +08:00
hiyouga	65cd8bdbdb	10x generate in ppo w/ zero3 ... https://github.com/huggingface/trl/pull/1483	2024-05-29 00:23:23 +08:00
hiyouga	c450ee87a3	improve KTO impl., replace datasets	2024-05-18 03:44:56 +08:00
hiyouga	308edbc426	rename package	2024-05-16 18:39:08 +08:00