Fix get_backend bug and add clear_device_cache function (#2857)

* added clear_device_cache

* set lambda: 0 for mps and cpu

src/accelerate/test_utils/testing.py

@@ -66,17 +66,17 @@ def get_backend():
     elif is_cuda_available():
         return "cuda", torch.cuda.device_count(), torch.cuda.memory_allocated
     elif is_mps_available(min_version="2.0"):
-        return "mps", 1, torch.mps.current_allocated_memory()
+        return "mps", 1, torch.mps.current_allocated_memory
     elif is_mps_available():
-        return "mps", 1, 0
+        return "mps", 1, lambda: 0
     elif is_mlu_available():
         return "mlu", torch.mlu.device_count(), torch.mlu.memory_allocated
     elif is_npu_available():
         return "npu", torch.npu.device_count(), torch.npu.memory_allocated
     elif is_xpu_available():
         return "xpu", torch.xpu.device_count(), torch.xpu.memory_allocated
     else:
-        return "cpu", 1, 0
+        return "cpu", 1, lambda: 0
 
 
 torch_device, device_count, memory_allocated_func = get_backend()

src/accelerate/utils/memory.py

@@ -26,6 +26,21 @@
 from .imports import is_mlu_available, is_mps_available, is_npu_available, is_xpu_available
 
 
+def clear_device_cache():
+    gc.collect()
+
+    if is_xpu_available():
+        torch.xpu.empty_cache()
+    elif is_mlu_available():
+        torch.mlu.empty_cache()
+    elif is_npu_available():
+        torch.npu.empty_cache()
+    elif is_mps_available(min_version="2.0"):
+        torch.mps.empty_cache()
+    else:
+        torch.cuda.empty_cache()
+
+
 def release_memory(*objects):
     """
     Releases memory from `objects` by setting them to `None` and calls `gc.collect()` and `torch.cuda.empty_cache()`.
@@ -52,17 +67,7 @@ def release_memory(*objects):
         objects = list(objects)
     for i in range(len(objects)):
         objects[i] = None
-    gc.collect()
-    if is_xpu_available():
-        torch.xpu.empty_cache()
-    elif is_mlu_available():
-        torch.mlu.empty_cache()
-    elif is_npu_available():
-        torch.npu.empty_cache()
-    elif is_mps_available(min_version="2.0"):
-        torch.mps.empty_cache()
-    else:
-        torch.cuda.empty_cache()
+    clear_device_cache()
     return objects
 
 
@@ -118,15 +123,7 @@ def find_executable_batch_size(function: callable = None, starting_batch_size: i
 
     def decorator(*args, **kwargs):
         nonlocal batch_size
-        gc.collect()
-        if is_xpu_available():
-            torch.xpu.empty_cache()
-        elif is_mlu_available():
-            torch.mlu.empty_cache()
-        elif is_npu_available():
-            torch.npu.empty_cache()
-        else:
-            torch.cuda.empty_cache()
+        clear_device_cache()
         params = list(inspect.signature(function).parameters.keys())
         # Guard against user error
         if len(params) < (len(args) + 1):
@@ -142,15 +139,7 @@ def decorator(*args, **kwargs):
                 return function(batch_size, *args, **kwargs)
             except Exception as e:
                 if should_reduce_batch_size(e):
-                    gc.collect()
-                    if is_xpu_available():
-                        torch.xpu.empty_cache()
-                    elif is_mlu_available():
-                        torch.mlu.empty_cache()
-                    elif is_npu_available():
-                        torch.npu.empty_cache()
-                    else:
-                        torch.cuda.empty_cache()
+                    clear_device_cache()
                     batch_size //= 2
                 else:
                     raise

src/accelerate/utils/modeling.py

@@ -39,6 +39,7 @@
     is_torch_xla_available,
     is_xpu_available,
 )
+from .memory import clear_device_cache
 from .offload import load_offloaded_weight, offload_weight, save_offload_index
 from .tqdm import is_tqdm_available, tqdm
 from .versions import compare_versions
@@ -456,14 +457,7 @@ def set_module_tensor_to_device(
                     module.weight = module.weight.cuda(device_index)
     # clean pre and post foward hook
     if device != "cpu":
-        if is_npu_available():
-            torch.npu.empty_cache()
-        elif is_mlu_available():
-            torch.mlu.empty_cache()
-        elif is_xpu_available():
-            torch.xpu.empty_cache()
-        else:
-            torch.cuda.empty_cache()
+        clear_device_cache()
 
     # When handling tied weights, we update tied_params_map to keep track of the tied weights that have already been allocated on the device in
     # order to avoid duplicating memory, see above.