NeoQuasar/Kronos-small
331.1m
1{2 "attn_dropout_p": 0.1,3 "d_model": 512,4 "ff_dim": 1024,5 "ffn_dropout_p": 0.25,6 "learn_te": true,7 "n_heads": 8,8 "n_layers": 8,9 "resid_dropout_p": 0.25,10 "s1_bits": 10,11 "s2_bits": 10,12 "token_dropout_p": 0.113}