[LoRA] device_map fix when loading LoRAs

docs/source/en/training/distributed_inference.md

@@ -237,3 +237,5 @@ with torch.no_grad():
 ```
 
 By selectively loading and unloading the models you need at a given stage and sharding the largest models across multiple GPUs, it is possible to run inference with large models on consumer GPUs.
+
+This workflow is also compatible with LoRAs via [`~loaders.StableDiffusionLoraLoaderMixin.load_lora_weights`]. However, only LoRAs without text encoder components are currently supported in this workflow.

src/diffusers/loaders/lora_base.py

@@ -211,12 +211,18 @@ def _optionally_disable_offloading(cls, _pipeline):
             tuple:
                 A tuple indicating if `is_model_cpu_offload` or `is_sequential_cpu_offload` is True.
         """
+        from ..pipelines.pipeline_loading_utils import model_has_device_map
+
         is_model_cpu_offload = False
         is_sequential_cpu_offload = False
 
         if _pipeline is not None and _pipeline.hf_device_map is None:
             for _, component in _pipeline.components.items():
-                if isinstance(component, nn.Module) and hasattr(component, "_hf_hook"):
+                if (
+                    isinstance(component, nn.Module)
+                    and hasattr(component, "_hf_hook")
+                    and not model_has_device_map(component)
+                ):
                     if not is_model_cpu_offload:
                         is_model_cpu_offload = isinstance(component._hf_hook, CpuOffload)
                     if not is_sequential_cpu_offload:

src/diffusers/loaders/unet.py

@@ -395,12 +395,18 @@ def _optionally_disable_offloading(cls, _pipeline):
             tuple:
                 A tuple indicating if `is_model_cpu_offload` or `is_sequential_cpu_offload` is True.
         """
+        from ..pipelines.pipeline_loading_utils import model_has_device_map
+
         is_model_cpu_offload = False
         is_sequential_cpu_offload = False
 
         if _pipeline is not None and _pipeline.hf_device_map is None:
             for _, component in _pipeline.components.items():
-                if isinstance(component, nn.Module) and hasattr(component, "_hf_hook"):
+                if (
+                    isinstance(component, nn.Module)
+                    and hasattr(component, "_hf_hook")
+                    and not model_has_device_map(component)
+                ):
                     if not is_model_cpu_offload:
                         is_model_cpu_offload = isinstance(component._hf_hook, CpuOffload)
                     if not is_sequential_cpu_offload:

src/diffusers/pipelines/pipeline_loading_utils.py

@@ -36,6 +36,7 @@
     deprecate,
     get_class_from_dynamic_module,
     is_accelerate_available,
+    is_accelerate_version,
     is_peft_available,
     is_transformers_available,
     logging,
@@ -947,3 +948,18 @@ def _get_ignore_patterns(
             )
 
     return ignore_patterns
+
+
+def model_has_device_map(model):
+    if not is_accelerate_available() or is_accelerate_version("<", "0.14.0"):
+        return False
+
+    # Check if the model has a device map that is not exclusively CPU
+    # `device_map` can only contain CPU when a model has sharded checkpoints.
+    # See here: https://github.com/huggingface/diffusers/blob/41e4779d988ead99e7acd78dc8e752de88777d0f/src/diffusers/models/modeling_utils.py#L883
+    device_map = getattr(model, "hf_device_map", None)
+    if device_map is not None:
+        unique_devices = set(device_map.values())
+        return len(unique_devices) > 1 or unique_devices != {"cpu"}
+
+    return False

src/diffusers/pipelines/pipeline_utils.py

@@ -85,6 +85,7 @@
     _update_init_kwargs_with_connected_pipeline,
     load_sub_model,
     maybe_raise_or_warn,
+    model_has_device_map,
     variant_compatible_siblings,
     warn_deprecated_model_variant,
 )
@@ -406,6 +407,16 @@ def module_is_offloaded(module):
 
             return hasattr(module, "_hf_hook") and isinstance(module._hf_hook, accelerate.hooks.CpuOffload)
 
+        # device-mapped modules should not go through any device placements.
+        device_mapped_components = [
+            key for key, component in self.components.items() if model_has_device_map(component)
+        ]
+        if device_mapped_components:
+            raise ValueError(
+                "The following pipeline components have been found to use a device map: "
+                f"{device_mapped_components}. This is incompatible with explicitly setting the device using `to()`."
+            )
+
         # .to("cuda") would raise an error if the pipeline is sequentially offloaded, so we raise our own to make it clearer
         pipeline_is_sequentially_offloaded = any(
             module_is_sequentially_offloaded(module) for _, module in self.components.items()
@@ -1002,6 +1013,16 @@ def enable_model_cpu_offload(self, gpu_id: Optional[int] = None, device: Union[t
                 The PyTorch device type of the accelerator that shall be used in inference. If not specified, it will
                 default to "cuda".
         """
+        # device-mapped modules should not go through any device placements.
+        device_mapped_components = [
+            key for key, component in self.components.items() if model_has_device_map(component)
+        ]
+        if device_mapped_components:
+            raise ValueError(
+                "The following pipeline components have been found to use a device map: "
+                f"{device_mapped_components}. This is incompatible with `enable_model_cpu_offload()`."
+            )
+
         is_pipeline_device_mapped = self.hf_device_map is not None and len(self.hf_device_map) > 1
         if is_pipeline_device_mapped:
             raise ValueError(
@@ -1104,6 +1125,16 @@ def enable_sequential_cpu_offload(self, gpu_id: Optional[int] = None, device: Un
                 The PyTorch device type of the accelerator that shall be used in inference. If not specified, it will
                 default to "cuda".
         """
+        # device-mapped modules should not go through any device placements.
+        device_mapped_components = [
+            key for key, component in self.components.items() if model_has_device_map(component)
+        ]
+        if device_mapped_components:
+            raise ValueError(
+                "The following pipeline components have been found to use a device map: "
+                f"{device_mapped_components}. This is incompatible with `enable_sequential_cpu_offload()`."
+            )
+
         if is_accelerate_available() and is_accelerate_version(">=", "0.14.0"):
             from accelerate import cpu_offload
         else:

tests/pipelines/audioldm2/test_audioldm2.py

@@ -506,9 +506,14 @@ def test_to_dtype(self):
         model_dtypes = {key: component.dtype for key, component in components.items() if hasattr(component, "dtype")}
         self.assertTrue(all(dtype == torch.float16 for dtype in model_dtypes.values()))
 
+    @unittest.skip("Test currently not supported.")
     def test_sequential_cpu_offload_forward_pass(self):
         pass
 
+    @unittest.skip("Test currently not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
 
 @nightly
 class AudioLDM2PipelineSlowTests(unittest.TestCase):

tests/pipelines/controlnet/test_controlnet.py

@@ -514,6 +514,18 @@ def test_inference_multiple_prompt_input(self):
 
         assert image.shape == (4, 64, 64, 3)
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 class StableDiffusionMultiControlNetOneModelPipelineFastTests(
     IPAdapterTesterMixin, PipelineTesterMixin, PipelineKarrasSchedulerTesterMixin, unittest.TestCase
@@ -697,6 +709,18 @@ def test_save_pretrained_raise_not_implemented_exception(self):
             except NotImplementedError:
                 pass
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 @slow
 @require_torch_gpu

tests/pipelines/controlnet/test_controlnet_img2img.py

@@ -389,6 +389,18 @@ def test_save_pretrained_raise_not_implemented_exception(self):
             except NotImplementedError:
                 pass
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 @slow
 @require_torch_gpu

tests/pipelines/controlnet/test_controlnet_inpaint.py

@@ -441,6 +441,18 @@ def test_save_pretrained_raise_not_implemented_exception(self):
             except NotImplementedError:
                 pass
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 @slow
 @require_torch_gpu

tests/pipelines/controlnet/test_controlnet_sdxl.py

@@ -683,6 +683,18 @@ def test_inference_batch_single_identical(self):
     def test_save_load_optional_components(self):
         return self._test_save_load_optional_components()
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 class StableDiffusionXLMultiControlNetOneModelPipelineFastTests(
     PipelineKarrasSchedulerTesterMixin, PipelineTesterMixin, SDXLOptionalComponentsTesterMixin, unittest.TestCase
@@ -887,6 +899,18 @@ def test_negative_conditions(self):
 
         self.assertTrue(np.abs(image_slice_without_neg_cond - image_slice_with_neg_cond).max() > 1e-2)
 
+    @unittest.skip("Test not supported.")
+    def test_calling_mco_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_to_raises_error_device_mapped_components(self):
+        pass
+
+    @unittest.skip("Test not supported.")
+    def test_calling_sco_raises_error_device_mapped_components(self):
+        pass
+
 
 @slow
 @require_torch_gpu