{
  "_name_or_path": "Ellio98/mistral-0.5B-base",
  "architectures": [
    "MistralForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "attn_logit_softcapping": 50.0,
  "bos_token_id": 1,
  "cache_implementation": "hybrid",
  "eos_token_id": 2,
  "final_logit_softcapping": 30.0,
  "head_dim": 192,
  "hidden_act": "silu",
  "hidden_activation": "gelu_pytorch_tanh",
  "hidden_size": 1536,
  "initializer_range": 0.02,
  "intermediate_size": 4096,
  "max_position_embeddings": 4096,
  "model_type": "mistral",
  "num_attention_heads": 8,
  "num_hidden_layers": 16,
  "num_key_value_heads": 4,
  "query_pre_attn_scalar": 256,
  "rms_norm_eps": 1e-05,
  "rope_theta": 10000.0,
  "sliding_window": 4096,
  "tie_word_embeddings": false,
  "torch_dtype": "float32",
  "transformers_version": "4.47.0",
  "use_cache": false,
  "vocab_size": 32768
}