[ONNX] Collate the external weights, speed up loading from the hub (#610)

Author: anton-l

scripts/convert_stable_diffusion_checkpoint_to_onnx.py

@@ -13,11 +13,14 @@
 # limitations under the License.
 
 import argparse
+import os
+import shutil
 from pathlib import Path
 
 import torch
 from torch.onnx import export
 
+import onnx
 from diffusers import StableDiffusionOnnxPipeline, StableDiffusionPipeline
 from diffusers.onnx_utils import OnnxRuntimeModel
 from packaging import version
@@ -92,10 +95,11 @@ def convert_models(model_path: str, output_path: str, opset: int):
     )
 
     # UNET
+    unet_path = output_path / "unet" / "model.onnx"
     onnx_export(
         pipeline.unet,
         model_args=(torch.randn(2, 4, 64, 64), torch.LongTensor([0, 1]), torch.randn(2, 77, 768), False),
-        output_path=output_path / "unet" / "model.onnx",
+        output_path=unet_path,
         ordered_input_names=["sample", "timestep", "encoder_hidden_states", "return_dict"],
         output_names=["out_sample"],  # has to be different from "sample" for correct tracing
         dynamic_axes={
@@ -106,6 +110,21 @@ def convert_models(model_path: str, output_path: str, opset: int):
         opset=opset,
         use_external_data_format=True,  # UNet is > 2GB, so the weights need to be split
     )
+    unet_model_path = str(unet_path.absolute().as_posix())
+    unet_dir = os.path.dirname(unet_model_path)
+    unet = onnx.load(unet_model_path)
+    # clean up existing tensor files
+    shutil.rmtree(unet_dir)
+    os.mkdir(unet_dir)
+    # collate external tensor files into one
+    onnx.save_model(
+        unet,
+        unet_model_path,
+        save_as_external_data=True,
+        all_tensors_to_one_file=True,
+        location="weights.pb",
+        convert_attribute=False,
+    )
 
     # VAE ENCODER
     vae_encoder = pipeline.vae

setup.py

@@ -90,8 +90,10 @@
     "isort>=5.5.4",
     "jax>=0.2.8,!=0.3.2,<=0.3.6",
     "jaxlib>=0.1.65,<=0.3.6",
-    "modelcards==0.1.4",
+    "modelcards>=0.1.4",
     "numpy",
+    "onnxruntime",
+    "onnxruntime-gpu",
     "pytest",
     "pytest-timeout",
     "pytest-xdist",
@@ -100,6 +102,7 @@
     "requests",
     "tensorboard",
     "torch>=1.4",
+    "torchvision",
     "transformers>=4.21.0",
 ]
 
@@ -171,10 +174,20 @@ def run(self):
 
 
 extras = {}
-extras["quality"] = ["black==22.8", "isort>=5.5.4", "flake8>=3.8.3", "hf-doc-builder"]
-extras["docs"] = ["hf-doc-builder"]
-extras["training"] = ["accelerate", "datasets", "tensorboard", "modelcards"]
-extras["test"] = ["datasets", "onnxruntime", "pytest", "pytest-timeout", "pytest-xdist", "scipy", "torchvision", "transformers"]
+extras["quality"] = deps_list("black", "isort", "flake8", "hf-doc-builder")
+extras["docs"] = deps_list("hf-doc-builder")
+extras["training"] = deps_list("accelerate", "datasets", "tensorboard", "modelcards")
+extras["test"] = deps_list(
+    "datasets",
+    "onnxruntime",
+    "onnxruntime-gpu",
+    "pytest",
+    "pytest-timeout",
+    "pytest-xdist",
+    "scipy",
+    "torchvision",
+    "transformers"
+)
 extras["torch"] = deps_list("torch")
 
 if os.name == "nt":  # windows

src/diffusers/dependency_versions_table.py

@@ -15,8 +15,10 @@
     "isort": "isort>=5.5.4",
     "jax": "jax>=0.2.8,!=0.3.2,<=0.3.6",
     "jaxlib": "jaxlib>=0.1.65,<=0.3.6",
-    "modelcards": "modelcards==0.1.4",
+    "modelcards": "modelcards>=0.1.4",
     "numpy": "numpy",
+    "onnxruntime": "onnxruntime",
+    "onnxruntime-gpu": "onnxruntime-gpu",
     "pytest": "pytest",
     "pytest-timeout": "pytest-timeout",
     "pytest-xdist": "pytest-xdist",
@@ -25,5 +27,6 @@
     "requests": "requests",
     "tensorboard": "tensorboard",
     "torch": "torch>=1.4",
+    "torchvision": "torchvision",
     "transformers": "transformers>=4.21.0",
 }

tests/test_pipelines.py

@@ -1373,12 +1373,9 @@ def test_stable_diffusion_inpaint_pipeline_k_lms(self):
 
     @slow
     def test_stable_diffusion_onnx(self):
-        from scripts.convert_stable_diffusion_checkpoint_to_onnx import convert_models
-
-        with tempfile.TemporaryDirectory() as tmpdirname:
-            convert_models("CompVis/stable-diffusion-v1-4", tmpdirname, opset=14)
-
-            sd_pipe = StableDiffusionOnnxPipeline.from_pretrained(tmpdirname, provider="CUDAExecutionProvider")
+        sd_pipe = StableDiffusionOnnxPipeline.from_pretrained(
+            "CompVis/stable-diffusion-v1-4", revision="onnx", provider="CUDAExecutionProvider", use_auth_token=True
+        )
 
         prompt = "A painting of a squirrel eating a burger"
         np.random.seed(0)