MIDI / config.json
sunsetsobserver's picture
End of training
d378a81 verified
{
"architectures": [
"MistralForCausalLM"
],
"attention_dropout": 0.0,
"bos_token_id": 2,
"eos_token_id": 3,
"hidden_act": "silu",
"hidden_size": 512,
"initializer_range": 0.02,
"intermediate_size": 2048,
"max_position_embeddings": 8192,
"model_type": "mistral",
"num_attention_heads": 8,
"num_hidden_layers": 8,
"num_key_value_heads": 4,
"pad_token_id": 0,
"rms_norm_eps": 1e-06,
"rope_theta": 10000.0,
"sliding_window": 256,
"tie_word_embeddings": false,
"torch_dtype": "float32",
"transformers_version": "4.37.2",
"use_cache": true,
"vocab_size": 10000
}