starcoder2-15b / config.json
lvwerra's picture
lvwerra HF staff
add dropout (#13)
acfb201 verified
raw
history blame
803 Bytes
{
"_name_or_path": "/aot/checkpoints/15b/config.json",
"architectures": [
"Starcoder2ForCausalLM"
],
"attention_dropout": 0.1,
"residual_dropout": 0.1,
"embedding_dropout": 0.1,
"bos_token_id": 0,
"eos_token_id": 0,
"hidden_act": "gelu_pytorch_tanh",
"hidden_size": 6144,
"initializer_range": 0.01275,
"intermediate_size": 24576,
"max_position_embeddings": 16384,
"mlp_type": "default",
"model_type": "starcoder2",
"norm_epsilon": 1e-05,
"norm_type": "layer_norm",
"num_attention_heads": 48,
"num_hidden_layers": 40,
"num_key_value_heads": 4,
"rope_theta": 100000,
"sliding_window": 4096,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.37.0.dev0",
"use_bias": true,
"use_cache": true,
"vocab_size": 49152
}