I have trained an LLM model using LoRa. Now I am trying to have access to the trained model for evaluation using the following code:
peft_trainer.train()
peft_model_path="C:\\Users\\Public\\Documents\\AI"
peft_trainer.model.save_pretrained(peft_model_path)
tokenizer.save_pretrained(peft_model_path)
instruct_model = AutoModelForSeq2SeqLM.from_pretrained("C:\\Users\\Public\\Documents\\AI", torch_dtype=torch.bfloat16)
It's giving me error:
OSError: ./peft-dialogue-summary-checkpoint-local does not appear to have a file named config.json
my folder in directory:
"C:\\Users\\Public\\Documents\\AI" has only managed to save files; adapter_config.json, adapter_model.bin,special_tokens_map.json,tokenizer.json,tokenizer_config.json.
How do I get the config.json file saved in the folder?
I tried to access my trained model. I was expecting the
instruct_model = AutoModelForSeq2SeqLM.from_pretrained("C:\\Users\\Public\\Documents\\AI", torch_dtype=torch.bfloat16)
to run without an error so that I should evaluate the model.