config.json

{
  "alibi": false,
  "apply_residual_connection_post_layernorm": false,
  "architectures": [
    "PheonixForCausalLM"
  ],
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_pheonix.PheonixConfig",
    "AutoModel": "modeling_pheonix.PheonixModel",
    "AutoModelForCausalLM": "modeling_pheonix.PheonixForCausalLM",
    "AutoModelForQuestionAnswering": "modeling_pheonix.PheonixForQuestionAnswering",
    "AutoModelForSequenceClassification": "modeling_pheonix.PheonixForSequenceClassification",
    "AutoModelForTokenClassification": "modeling_pheonix.PheonixForTokenClassification"
  },
  "bias": false,
  "bos_token_id": 11,
  "dtype": "float32",
  "eos_token_id": 11,
  "hidden_dropout": 0.0,
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "layer_norm_epsilon": 1e-05,
  "model_type": "Pheonix",
  "multi_query": true,
  "new_decoder_architecture": false,
  "num_attention_heads": 64,
  "num_hidden_layers": 24,
  "num_kv_heads": 64,
  "parallel_attn": true,
  "transformers_version": "5.0.0",
  "use_cache": true,
  "vocab_size": 65024
}