specify low_cpu_mem_usage=True to halve loading requirements

run_test.py

@@ -17,7 +17,7 @@ os.environ['CUDA_LAUNCH_BLOCKING'] = '1'
 torch.manual_seed(0)
 random.seed(0)
 np.random.seed(0)
-import accelerate, safetensors.torch, transformers, torch, tqdm
+import accelerate, safetensors.torch, transformers, tqdm
 
 config = transformers.AutoConfig.from_pretrained(model_id, revision=revision, trust_remote_code=True)
 tokenizer = transformers.AutoTokenizer.from_pretrained(model_id, revision=revision, trust_remote_code=True)
@@ -25,7 +25,7 @@ Model = transformers.AutoModelForCausalLM
 if config.model_type == 'deepseek_v3':
     #Model = transformers.DeepseekV3ForCausalLM
     pass
-model = Model.from_pretrained(model_id, revision=revision, trust_remote_code=True, torch_dtype=torch.float64, use_safetensors=True)
+model = Model.from_pretrained(model_id, revision=revision, trust_remote_code=True, torch_dtype=torch.float64, use_safetensors=True, low_cpu_mem_usage=True)
 if config.model_type == 'deepseek_v3':
     model._supports_cache_class = False
 #model = accelerate.cpu_offload(model, 'cuda:0', offload_buffers=True)