6 Commits

Author	SHA1	Message	Date
Kirk Lin	e31720d663	Merge 837b43a504 into 190d9515d0	2025-10-15 18:47:38 +02:00
Andrej Karpathy	190d9515d0	dont evaluate the sampling evals during SFT they are too slow. keep the multiple choice evals. delete unused imports	2025-10-15 16:42:23 +00:00
Andrej Karpathy	b8076dd367	fix bug in learning rate multiplier, it was ramping up instead of ramping down. see more in Issue #68. also add --dry_run option useful for experimentation	2025-10-15 16:35:04 +00:00
Kirk Lin	662ff7eb7a	feat: dynamic dtype selection	2025-10-14 12:22:57 +08:00
Kirk Lin	447567634c	feat: cross-platform support for CPU and GPU environments	2025-10-14 12:11:37 +08:00
karpathy	3a5e0bc50b	initial commit	2025-10-13 06:49:24 -07:00