Is this the right adapter_config.json?

#1
by TheBloke - opened

Hi

Thanks for uploading this interesting new LoRA.

You didn't include adapter_config.json which makes it hard to load the LoRA.

Reading the description and comments, I think this is correct? Could you confirm, or push the original adapter_config.json?

Many thanks

{
  "base_model_name_or_path": "decapoda-research/llama-7b-hf",
  "bias": "none",
  "enable_lora": null,
  "fan_in_fan_out": false,
  "inference_mode": true,
  "lora_alpha": 16,
  "lora_dropout": 0.05,
  "merge_weights": false,
  "modules_to_save": null,
  "peft_type": "LORA",
  "r": 16,
  "target_modules": [
    "q_proj",
    "k_proj",
    "v_proj",
    "o_proj"
  ],
  "task_type": "CAUSAL_LM"
}

Hi, here is the adapter_config.json we used:

{
  "base_model_name_or_path": "llama-65b-hf",
  "bias": "none",
  "enable_lora": null,
  "fan_in_fan_out": false,
  "inference_mode": true,
  "lora_alpha": 16,
  "lora_dropout": 0.05,
  "merge_weights": false,
  "modules_to_save": null,
  "peft_type": "LORA",
  "r": 16,
  "target_modules": [
    "q_proj",
    "k_proj",
    "v_proj",
    "o_proj"
  ],
  "task_type": "CAUSAL_LM"
}
zhiqings changed discussion status to closed

Thank you very much

Sign up or log in to comment