photo-classifier / config.json
winglian's picture
Training in progress, epoch 1
057658b
raw
history blame
791 Bytes
{
"_name_or_path": "winglian/basilisk-4b",
"architectures": [
"LlamaForSequenceClassification"
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 4096,
"id2label": {
"0": "NO",
"1": "YES"
},
"initializer_range": 0.02,
"intermediate_size": 11008,
"label2id": {
"NO": 0,
"YES": 1
},
"max_position_embeddings": 4096,
"model_type": "llama",
"num_attention_heads": 32,
"num_hidden_layers": 16,
"num_key_value_heads": 32,
"pretraining_tp": 1,
"problem_type": "single_label_classification",
"rms_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 10000.0,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.34.0.dev0",
"use_cache": true,
"vocab_size": 32002
}