56 Commits

Author	SHA1	Message	Date
Goekdeniz-Guelmez	1f89453295	eos token return fix	2025-03-05 14:00:51 +01:00
Goekdeniz-Guelmez	2bde97fe13	minor speed improvement	2025-03-05 13:55:24 +01:00
Goekdeniz-Guelmez	3dfb21267b	updates	2025-03-05 12:59:41 +01:00
Goekdeniz-Guelmez	132225a018	updates	2025-03-01 22:23:33 +01:00
Goekdeniz-Guelmez	925e11439b	updates	2025-02-28 22:07:24 +01:00
Goekdeniz-Guelmez	15d53279ae	batching fix	2025-02-28 16:02:40 +01:00
Goekdeniz-Guelmez	fab2dc2688	smoll fix	2025-02-26 15:21:57 +01:00
Goekdeniz-Guelmez	53185c7f3d	last update, gn	2025-02-24 22:20:07 +01:00
Goekdeniz-Guelmez	e4eac9c97b	adding custom system message integration in dataset, more opimizations (generates now faster, while same RAM usage), fix for the identical generatrions, seperated the reward functions into a seperate file.	2025-02-24 20:49:22 +01:00
Goekdeniz-Guelmez	9705ed908e	fix wrong generation in train	2025-02-22 17:21:08 +01:00
Goekdeniz-Guelmez	d9c4c6e60c	clean up and readding temperature argument	2025-02-22 02:34:56 +01:00
Goekdeniz-Guelmez	d653371e3d	nits	2025-02-22 02:12:02 +01:00
Goekdeniz-Guelmez	235348c211	generation speed improvement in training too from 3 t/s to 15 t/s	2025-02-22 02:03:01 +01:00
Goekdeniz-Guelmez	79de353530	nits	2025-02-22 01:05:58 +01:00
Goekdeniz-Guelmez	c51b0a2715	fix	2025-02-22 00:21:47 +01:00
Goekdeniz-Guelmez	710bc1490e	training mode working too got from 2 toks/sec to 30 toks/sec with raw 1.5B model	2025-02-21 22:42:15 +01:00
Goekdeniz-Guelmez	6086137131	Huge speed improvement in validation mode.	2025-02-21 22:08:49 +01:00
Goekdeniz-Guelmez	2f20107d9b	little faster generation + prints ot a examplke generatino in validation mode, more optimization in trianing function	2025-02-21 16:02:27 +01:00
Goekdeniz-Guelmez	541f0be937	fix generation cutoff in evaluation	2025-02-17 14:39:38 +01:00
Goekdeniz-Guelmez	6a6bd53e43	removing print and switching some variables in the math	2025-02-15 15:38:51 +01:00
Goekdeniz-Guelmez	5ec4790656	removing comments + adding temperature + reward weighting	2025-02-15 15:29:22 +01:00
Goekdeniz-Guelmez	e33d9d509b	updates	2025-02-12 11:07:53 +01:00
Goekdeniz-Guelmez	5aeefc8c47	update new iterade batches function + nits	2025-02-12 08:57:26 +01:00
Goekdeniz-Guelmez	e80bf95182	fix	2025-02-11 09:26:43 +01:00
Goekdeniz-Guelmez	e96afe9e9f	updates	2025-02-11 09:09:28 +01:00
Goekdeniz-Guelmez	88ca747e9e	nits	2025-02-10 19:46:19 +01:00
Goekdeniz-Guelmez	b7bc811507	nits	2025-02-10 19:45:19 +01:00
Goekdeniz-Guelmez	e5aa2c3b5d	nits	2025-02-10 17:51:14 +01:00
Goekdeniz-Guelmez	f88e897019	removing helper functions	2025-02-10 16:07:28 +01:00
Goekdeniz-Guelmez	d9da35f458	nits	2025-02-10 10:52:32 +01:00
Goekdeniz-Guelmez	00712522ba	rebase loss calculation	2025-02-09 17:13:05 +01:00
Goekdeniz-Guelmez	a527cdb39b	fix: prevent gradients from flowing through the reference model's logits	2025-02-09 17:02:58 +01:00
Goekdeniz-Guelmez	54179901b5	fix	2025-02-09 15:41:47 +01:00
Goekdeniz-Guelmez	9ba6146a76	fix	2025-02-09 14:32:50 +01:00
Goekdeniz-Guelmez	bcfa55d882	updates	2025-02-05 15:02:12 +01:00
Goekdeniz-Guelmez	0a19522ec4	updates	2025-02-05 14:38:09 +01:00
Goekdeniz-Guelmez	35a2d99cf9	smoll fix	2025-02-05 11:30:21 +01:00
Goekdeniz-Guelmez	a33cad84b4	udpates	2025-02-05 09:48:00 +01:00
Goekdeniz-Guelmez	2a8e6f6e44	udpate	2025-02-05 08:47:03 +01:00
Goekdeniz-Guelmez	0a09a93454	fix cache handling	2025-02-05 08:44:06 +01:00
Goekdeniz-Guelmez	7173840283	first succesfull training run	2025-02-04 09:18:45 +01:00
Goekdeniz-Guelmez	ca32424043	updates	2025-02-03 21:57:26 +01:00
Goekdeniz-Guelmez	54e295ea80	fix name funcs	2025-02-03 19:56:11 +01:00
Goekdeniz-Guelmez	06f9c29c94	print func name	2025-02-03 19:47:40 +01:00
Goekdeniz-Guelmez	40bca770ae	fixes	2025-02-03 19:43:49 +01:00
Goekdeniz-Guelmez	05d921b788	optims	2025-02-03 19:37:05 +01:00
Goekdeniz-Guelmez	1d9e4802f0	first working prototype, will try training out at home	2025-02-03 12:05:29 +01:00
Goekdeniz-Guelmez	23d75cd7ad	starting fist training test run	2025-02-03 10:08:28 +01:00
Goekdeniz-Guelmez	a3ed632422	dataset wrapper done	2025-02-03 09:13:17 +01:00
Goekdeniz-Guelmez	d034ca369e	adding function for R1	2025-02-03 08:26:42 +01:00