config.json

{
  "_name_or_path": "results/pretrain/160M_pds",
  "architectures": [
    "MistralForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "head_dim": 64,
  "hidden_act": "silu",
  "hidden_size": 768,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "is_model_parallel": false,
  "max_position_embeddings": 32768,
  "model_type": "mistral",
  "num_attention_heads": 12,
  "num_hidden_layers": 12,
  "num_key_value_heads": 12,
  "rms_norm_eps": 1e-05,
  "rope_theta": 10000.0,
  "sliding_window": 4096,
  "tie_word_embeddings": false,
  "torch_dtype": "float32",
  "transformers_version": "4.46.0.dev0",
  "use_cache": true,
  "use_memory_efficient_attention": true,
  "vocab_size": 32000
}