maximeseince's picture
Training in progress, step 1
948d1a3 verified
raw
history blame
197 Bytes
{"d_model": 1024, "n_layer": 16, "vocab_size": 38, "ssm_cfg": {}, "rms_norm": true, "residual_in_fp32": false, "fused_add_norm": true, "pad_vocab_size_multiple": 8, "max_position_embeddings": 2048}