Enable BitsAndBytes quantization in MPS (#13915)

Fix BitsAndBytes quantization in MPS

Co-authored-by: Sayak Paul <spsayakpaul@gmail.com>

Author: LucasSte

src/diffusers/quantizers/bitsandbytes/bnb_quantizer.py

@@ -61,7 +61,7 @@ def __init__(self, quantization_config, **kwargs):
             self.modules_to_not_convert = self.quantization_config.llm_int8_skip_modules
 
     def validate_environment(self, *args, **kwargs):
-        if not (torch.cuda.is_available() or torch.xpu.is_available()):
+        if not (torch.cuda.is_available() or torch.xpu.is_available() or torch.mps.is_available()):
             raise RuntimeError("No GPU found. A GPU is needed for quantization.")
         if not is_accelerate_available() or is_accelerate_version("<", "0.26.0"):
             raise ImportError(
@@ -240,6 +240,8 @@ def update_device_map(self, device_map):
         if device_map is None:
             if torch.xpu.is_available():
                 current_device = f"xpu:{torch.xpu.current_device()}"
+            elif torch.mps.is_available():
+                current_device = "mps"
             else:
                 current_device = f"cuda:{torch.cuda.current_device()}"
             device_map = {"": current_device}
@@ -411,6 +413,8 @@ def update_device_map(self, device_map):
         if device_map is None:
             if torch.xpu.is_available():
                 current_device = f"xpu:{torch.xpu.current_device()}"
+            elif torch.mps.is_available():
+                current_device = "mps"
             else:
                 current_device = f"cuda:{torch.cuda.current_device()}"
             device_map = {"": current_device}