device_map

README.md

@@ -119,6 +119,9 @@ m.save_pretrained("nyc-savvy")
 You might wonder if the model successfully learned anything about NYC or is the same old LLaMa2. With your prompt not adding clues, try this from the `pefttester.py` script in this repo:
 
 ```python
+m = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")
+tok = LlamaTokenizer.from_pretrained(model_name)
+
 messages = "A chat between a curious human and an assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.\n"
 messages += "### Human: What museums should I visit? - My kids are aged 12 and 5"
 messages += "### Assistant: "