llm-hw1 / config.json
mcnckc's picture
Push model using huggingface_hub.
225d73b verified
raw
history blame contribute delete
180 Bytes
{
"dropout": 0.1,
"hidden_dim": 96,
"intermediate_dim": 256,
"max_seq_len": 128,
"n_head": 4,
"n_kv_head": 2,
"n_layer": 3,
"use_RoPE": true,
"vocab_size": 1024
}