16 Commits

Author SHA1 Message Date
k
0537a5df64 changed chat dataset. 2026-01-09 17:30:34 -05:00
k
c78a31362a set to gpt2 hyprs 2026-01-09 12:45:01 -05:00
k
496916f428 added fine-tuning 2026-01-07 13:01:06 -05:00
k
121640bab6 updated hypr for my gpu 2026-01-07 12:59:44 -05:00
k
6f037c4a9a Quick training script 2026-01-07 02:14:09 -05:00
k
7f25dff1d1 Fix errors 2026-01-07 02:13:08 -05:00
k
007c96e91b Simple log functions 2026-01-07 01:25:47 -05:00
k
6daa8ec46c Added code to generate training batches 2026-01-07 01:15:18 -05:00
k
229c564811 CosineAnnealing with optimizer Group 2026-01-07 00:26:04 -05:00
k
478010c8cc added Positional encodeings 2026-01-06 21:38:12 -05:00
k
3b590b3ce7 added dropout to ffn 2026-01-06 21:26:51 -05:00
k
957aad2239 Implimented Transformer(decode only) 2026-01-06 21:26:24 -05:00
k
23f62c7e64 Implimented TransformerBlock 2026-01-06 19:53:37 -05:00
k
77aa0de0eb Implimented MultiHeadAttention 2026-01-06 19:41:12 -05:00
k
c4e5e332ba fixed cast in ffn 2026-01-06 19:40:45 -05:00
k
d6b9f45fcc Implimented Feed Forward Netwok 2026-01-06 18:31:04 -05:00