X-EcoMLA-1B8B-fixed-kv64-DPO / MLA_config.json
Mingyuyang-1's picture
Upload folder using huggingface_hub
aca3f31
raw
history blame contribute delete
212 Bytes
{
"d_model": 2048,
"rms_norm_eps": 1e-05,
"vocab_size": null,
"d_inner": 2048,
"d_xb": 512,
"intermediate_size": 8192,
"hidden_act": "silu",
"n_layer": 16,
"attn_layers": []
}