EddieChen372's picture
Training in progress step 10000
c8233be
{
"_name_or_path": "/checkpoint/dpf/models/cm-1.3B-hf",
"activation_dropout": 0.0,
"activation_function": "gelu",
"architectures": [
"XGLMForCausalLM"
],
"attention_dropout": 0.1,
"attention_heads": 32,
"bos_token_id": 0,
"d_model": 1600,
"decoder_start_token_id": 2,
"dropout": 0.1,
"eos_token_id": 0,
"ffn_dim": 8192,
"init_std": 0.02,
"layerdrop": 0.0,
"max_position_embeddings": 512,
"model_type": "xglm",
"num_layers": 12,
"pad_token_id": 0,
"scale_embedding": true,
"torch_dtype": "float32",
"transformers_version": "4.20.1",
"use_cache": true,
"vocab_size": 52003
}