checkpoint-700/config.json

{
  "_name_or_path": "lmsys/vicuna-13b-v1.5",
  "agent_encoder_dim": 1024,
  "agent_num_encoders": 15,
  "agent_projector_lr": null,
  "agent_state_tokens": 64,
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "hidden_act": "silu",
  "hidden_size": 5120,
  "initializer_range": 0.02,
  "intermediate_size": 13824,
  "llm_state_tokens": 16,
  "max_length": 4096,
  "max_position_embeddings": 4096,
  "mlp_bias": false,
  "model_type": "arjun_llama",
  "num_attention_heads": 40,
  "num_hidden_layers": 40,
  "num_key_value_heads": 40,
  "pad_token_id": 0,
  "pretraining_tp": 1,
  "projection_layer": "encoder[\\d]+/ln2",
  "projector_dropout": 0.1,
  "projector_hidden_depth": 3,
  "projector_hidden_dim": 2048,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "tie_word_embeddings": false,
  "tokenizer_model_max_length": 375,
  "tokenizer_padding_side": "right",
  "torch_dtype": "float16",
  "transformers_version": "4.43.1",
  "tune_agent_adapter": true,
  "use_cache": false,
  "use_state_patch_token": true,
  "use_state_start_end": false,
  "vocab_size": 32001
}