Configuration.json

{
  "architectures": ["MistralForCausalLM"],
  "model_type": "mistral",
  "vocab_size": 32000,
  "hidden_size": 4096,
  "num_hidden_layers": 32,
  "num_attention_heads": 32,
  "intermediate_size": 14336,
  "max_position_embeddings": 32768,
  "initializer_range": 0.02,
  "rope_theta": 1000000.0
}