MTSAIR

/

MultiVerse_70B

like 39

Follow

MTSAIR 25

Hi there, just want to confirm the base model this was trained from? The config.json file has a discrepancy:

{
  "_name_or_path": "",
  "architectures": [
    "LlamaForCausalLM".   <--- in Qwen1.5/2 models, this is "Qwen2ForCausalLM"
  ],
  "attention_bias": true,
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 151643,
  "hidden_act": "silu",
  "hidden_size": 8192,
  "initializer_range": 0.02,
  "intermediate_size": 24576,
  "max_position_embeddings": 32768,
  "model_type": "llama",    <--- in Qwen1.5/2 models, this is "qwen2"
  "num_attention_heads": 64,
  "num_hidden_layers": 80,
  "num_key_value_heads": 64,
  "pad_token_id": 151643,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000,
  "seq_length": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.38.2",
  "use_cache": true,
  "vocab_size": 152064
}