LLaMA-MLM-Large / config.json
hyen's picture
Upload LlamaForMLM
a72fd31 verified
raw
history blame contribute delete
769 Bytes
{
"_name_or_path": "/scratch/gpfs/hyen/p-cross-att/output/llama-large-mlm/checkpoint-100000",
"architectures": [
"LlamaForMLM"
],
"attention_bias": false,
"bos_token_id": 1,
"do_cross_attention": false,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 1024,
"initializer_range": 0.02,
"intermediate_size": 4096,
"is_encoder": true,
"max_position_embeddings": 512,
"model_type": "llama",
"num_attention_heads": 16,
"num_hidden_layers": 24,
"num_key_value_heads": 16,
"pad_token_id": 0,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": true,
"torch_dtype": "float32",
"transformers_version": "4.34.1",
"use_cache": true,
"vocab_size": 32001
}