diff --git a/train.py b/train.py index 1406b83..6991b6c 100644 --- a/train.py +++ b/train.py @@ -41,7 +41,7 @@ def train(config_path): model = AutoModelForCausalLM.from_pretrained( config["base_model"], torch_dtype=torch.float16, - device_map="auto", + # device_map="auto", trust_remote_code=True, ) print("✓ Success: 4-bit AS-IS") @@ -81,7 +81,7 @@ def train(config_path): model = AutoModelForCausalLM.from_pretrained( config["base_model"], torch_dtype=torch.bfloat16, - device_map="auto", + # device_map="auto", low_cpu_mem_usage=True, trust_remote_code=True, )