[Dtype] Align dtype casting behavior with Transformers and Accelerate (#1725)

* [Dtype] Align automatic dtype

* up

* up

* fix

* re-add accelerate

Author: patrickvonplaten

.github/workflows/nightly_tests.yml

@@ -62,6 +62,7 @@ jobs:
         run: |
           python -m pip install -e .[quality,test]
           python -m pip install -U git+https://github.com/huggingface/transformers
+          python -m pip install git+https://github.com/huggingface/accelerate
 
       - name: Environment
         run: |
@@ -134,6 +135,7 @@ jobs:
           ${CONDA_RUN} python -m pip install --upgrade pip
           ${CONDA_RUN} python -m pip install -e .[quality,test]
           ${CONDA_RUN} python -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
+          ${CONDA_RUN} python -m pip install git+https://github.com/huggingface/accelerate
 
       - name: Environment
         shell: arch -arch arm64 bash {0}
@@ -157,4 +159,4 @@ jobs:
         uses: actions/upload-artifact@v2
         with:
           name: torch_mps_test_reports
-          path: reports
+          path: reports

.github/workflows/pr_tests.yml

@@ -60,6 +60,7 @@ jobs:
         apt-get update && apt-get install libsndfile1-dev -y
         python -m pip install -e .[quality,test]
         python -m pip install -U git+https://github.com/huggingface/transformers
+        python -m pip install git+https://github.com/huggingface/accelerate
 
     - name: Environment
       run: |
@@ -126,6 +127,7 @@ jobs:
         ${CONDA_RUN} python -m pip install --upgrade pip
         ${CONDA_RUN} python -m pip install -e .[quality,test]
         ${CONDA_RUN} python -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
+        ${CONDA_RUN} python -m pip install git+https://github.com/huggingface/accelerate
         ${CONDA_RUN} python -m pip install -U git+https://github.com/huggingface/transformers
 
     - name: Environment

.github/workflows/push_tests.yml

@@ -62,6 +62,7 @@ jobs:
       run: |
         python -m pip install -e .[quality,test]
         python -m pip install -U git+https://github.com/huggingface/transformers
+        python -m pip install git+https://github.com/huggingface/accelerate
 
     - name: Environment
       run: |
@@ -130,6 +131,7 @@ jobs:
     - name: Install dependencies
       run: |
         python -m pip install -e .[quality,test,training]
+        python -m pip install git+https://github.com/huggingface/accelerate
         python -m pip install -U git+https://github.com/huggingface/transformers
 
     - name: Environment
@@ -151,4 +153,4 @@ jobs:
       uses: actions/upload-artifact@v2
       with:
         name: examples_test_reports
-        path: reports
+        path: reports

src/diffusers/modeling_utils.py

@@ -14,6 +14,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import inspect
 import os
 from functools import partial
 from typing import Callable, List, Optional, Tuple, Union
@@ -489,11 +490,15 @@ def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.P
                 state_dict = load_state_dict(model_file)
                 # move the parms from meta device to cpu
                 for param_name, param in state_dict.items():
-                    set_module_tensor_to_device(model, param_name, param_device, value=param)
+                    accepts_dtype = "dtype" in set(inspect.signature(set_module_tensor_to_device).parameters.keys())
+                    if accepts_dtype:
+                        set_module_tensor_to_device(model, param_name, param_device, value=param, dtype=torch_dtype)
+                    else:
+                        set_module_tensor_to_device(model, param_name, param_device, value=param)
             else:  # else let accelerate handle loading and dispatching.
                 # Load weights and dispatch according to the device_map
                 # by deafult the device_map is None and the weights are loaded on the CPU
-                accelerate.load_checkpoint_and_dispatch(model, model_file, device_map)
+                accelerate.load_checkpoint_and_dispatch(model, model_file, device_map, dtype=torch_dtype)
 
             loading_info = {
                 "missing_keys": [],
@@ -519,20 +524,6 @@ def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.P
             model = cls.from_config(config, **unused_kwargs)
 
             state_dict = load_state_dict(model_file)
-            dtype = set(v.dtype for v in state_dict.values())
-
-            if len(dtype) > 1 and torch.float32 not in dtype:
-                raise ValueError(
-                    f"The weights of the model file {model_file} have a mixture of incompatible dtypes {dtype}. Please"
-                    f" make sure that {model_file} weights have only one dtype."
-                )
-            elif len(dtype) > 1 and torch.float32 in dtype:
-                dtype = torch.float32
-            else:
-                dtype = dtype.pop()
-
-            # move model to correct dtype
-            model = model.to(dtype)
 
             model, missing_keys, unexpected_keys, mismatched_keys, error_msgs = cls._load_pretrained_model(
                 model,

tests/test_modeling_common.py

@@ -70,9 +70,9 @@ def test_from_save_pretrained_dtype(self):
             with tempfile.TemporaryDirectory() as tmpdirname:
                 model.to(dtype)
                 model.save_pretrained(tmpdirname)
-                new_model = self.model_class.from_pretrained(tmpdirname, low_cpu_mem_usage=True)
+                new_model = self.model_class.from_pretrained(tmpdirname, low_cpu_mem_usage=True, torch_dtype=dtype)
                 assert new_model.dtype == dtype
-                new_model = self.model_class.from_pretrained(tmpdirname, low_cpu_mem_usage=False)
+                new_model = self.model_class.from_pretrained(tmpdirname, low_cpu_mem_usage=False, torch_dtype=dtype)
                 assert new_model.dtype == dtype
 
     def test_determinism(self):