MambaByte_Code / config.json
JunxiongWang's picture
Upload 2 files
d8412cf verified
raw
history blame contribute delete
No virus
213 Bytes
{"d_model": 1024, "n_layer": 56, "vocab_size": 256, "ssm_cfg": {"expand": 2}, "rms_norm": true, "residual_in_fp32": true, "fused_add_norm": true, "pad_vocab_size_multiple": 8, "tie_embeddings": false, "pad_id": 0}