TransNormerLLM-1B / config.json
OpenNLPLab's picture
Publish 1B Model
92d0fa0
{
"architectures": [
"TransnormerForCausalLM"
],
"auto_map": {
"AutoConfig": "configuration_transnormer.TransnormerConfig",
"AutoModelForCausalLM": "modeling_transnormer.TransnormerForCausalLM"
},
"pad_token_id": 0,
"bos_token_id": 1,
"eos_token_id": 2,
"vocab_size": 64000,
"use_cache": true,
"init_std": 0.02,
"decoder_embed_dim": 2048,
"decoder_layers": 16,
"decoder_attention_heads": 16,
"no_scale_embedding": false,
"add_bos_token": false,
"norm_type": "simplermsnorm",
"linear_use_lrpe_list": [
1,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
],
"hidden_dim": 2048,
"linear_act_fun": "swish",
"glu_dim": 5632,
"bias": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.32.0"
}