Remove modelcards dependency (huggingface#2050)

Wauplin · web-flow · commit bcb476797ccb · 2023-01-20T16:39:42.000+01:00
* Switch to huggingface_hub.ModelCard

* Remove modelcards dependency in favor of Jinja2
diff --git a/docker/diffusers-flax-cpu/Dockerfile b/docker/diffusers-flax-cpu/Dockerfile
@@ -34,8 +34,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/docker/diffusers-flax-tpu/Dockerfile b/docker/diffusers-flax-tpu/Dockerfile
@@ -36,8 +36,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \        
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/docker/diffusers-onnxruntime-cpu/Dockerfile b/docker/diffusers-onnxruntime-cpu/Dockerfile
@@ -34,8 +34,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/docker/diffusers-onnxruntime-cuda/Dockerfile b/docker/diffusers-onnxruntime-cuda/Dockerfile
@@ -34,8 +34,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/docker/diffusers-pytorch-cpu/Dockerfile b/docker/diffusers-pytorch-cpu/Dockerfile
@@ -33,8 +33,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/docker/diffusers-pytorch-cuda/Dockerfile b/docker/diffusers-pytorch-cuda/Dockerfile
@@ -33,8 +33,8 @@ RUN python3 -m pip install --no-cache-dir --upgrade pip && \
         datasets \
         hf-doc-builder \
         huggingface-hub \
+        Jinja2 \
         librosa \
-        modelcards \
         numpy \
         scipy \
         tensorboard \
diff --git a/examples/dreambooth/requirements.txt b/examples/dreambooth/requirements.txt
@@ -3,4 +3,4 @@ torchvision
 transformers>=4.25.1
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/dreambooth/requirements_flax.txt b/examples/dreambooth/requirements_flax.txt
@@ -5,4 +5,4 @@ torch
 torchvision
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/research_projects/colossalai/requirement.txt b/examples/research_projects/colossalai/requirement.txt
@@ -3,5 +3,5 @@ torch
 torchvision
 ftfy
 tensorboard
-modelcards
+Jinja2
 transformers
diff --git a/examples/research_projects/dreambooth_inpaint/requirements.txt b/examples/research_projects/dreambooth_inpaint/requirements.txt
@@ -4,4 +4,4 @@ torchvision
 transformers>=4.21.0
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/research_projects/intel_opts/textual_inversion/requirements.txt b/examples/research_projects/intel_opts/textual_inversion/requirements.txt
@@ -3,5 +3,5 @@ torchvision
 transformers>=4.21.0
 ftfy
 tensorboard
-modelcards
+Jinja2
 intel_extension_for_pytorch>=1.13
diff --git a/examples/research_projects/multi_subject_dreambooth/requirements.txt b/examples/research_projects/multi_subject_dreambooth/requirements.txt
@@ -3,4 +3,4 @@ torchvision
 transformers>=4.25.1
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/text_to_image/requirements.txt b/examples/text_to_image/requirements.txt
@@ -4,4 +4,4 @@ transformers>=4.25.1
 datasets
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/text_to_image/requirements_flax.txt b/examples/text_to_image/requirements_flax.txt
@@ -6,4 +6,4 @@ torch
 torchvision
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/textual_inversion/requirements.txt b/examples/textual_inversion/requirements.txt
@@ -3,4 +3,4 @@ torchvision
 transformers>=4.25.1
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/examples/textual_inversion/requirements_flax.txt b/examples/textual_inversion/requirements_flax.txt
@@ -5,4 +5,4 @@ torch
 torchvision
 ftfy
 tensorboard
-modelcards
+Jinja2
diff --git a/setup.py b/setup.py
@@ -91,9 +91,9 @@
     "isort>=5.5.4",
     "jax>=0.2.8,!=0.3.2",
     "jaxlib>=0.1.65",
+    "Jinja2",
     "k-diffusion>=0.0.12",
     "librosa",
-    "modelcards>=0.1.4",
     "numpy",
     "parameterized",
     "pytest",
@@ -180,9 +180,10 @@ def run(self):
 extras = {}
 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["training"] = deps_list("accelerate", "datasets", "tensorboard", "Jinja2")
 extras["test"] = deps_list(
     "datasets",
+    "Jinja2",
     "k-diffusion",
     "librosa",
     "parameterized",
diff --git a/src/diffusers/dependency_versions_table.py b/src/diffusers/dependency_versions_table.py
@@ -15,9 +15,9 @@
     "isort": "isort>=5.5.4",
     "jax": "jax>=0.2.8,!=0.3.2",
     "jaxlib": "jaxlib>=0.1.65",
+    "Jinja2": "Jinja2",
     "k-diffusion": "k-diffusion>=0.0.12",
     "librosa": "librosa",
-    "modelcards": "modelcards>=0.1.4",
     "numpy": "numpy",
     "parameterized": "parameterized",
     "pytest": "pytest",
diff --git a/src/diffusers/utils/__init__.py b/src/diffusers/utils/__init__.py
@@ -48,7 +48,6 @@
     is_k_diffusion_available,
     is_k_diffusion_version,
     is_librosa_available,
-    is_modelcards_available,
     is_onnx_available,
     is_safetensors_available,
     is_scipy_available,
diff --git a/src/diffusers/utils/hub_utils.py b/src/diffusers/utils/hub_utils.py
@@ -20,7 +20,8 @@
 from typing import Dict, Optional, Union
 from uuid import uuid4
 
-from huggingface_hub import HfFolder, whoami
+from huggingface_hub import HfFolder, ModelCard, ModelCardData, whoami
+from huggingface_hub.utils import is_jinja_available
 
 from .. import __version__
 from .constants import HUGGINGFACE_CO_RESOLVE_ENDPOINT
@@ -31,21 +32,16 @@
     _onnxruntime_version,
     _torch_version,
     is_flax_available,
-    is_modelcards_available,
     is_onnx_available,
     is_torch_available,
 )
 from .logging import get_logger
 
 
-if is_modelcards_available():
-    from modelcards import CardData, ModelCard
-
-
 logger = get_logger(__name__)
 
 
-MODEL_CARD_TEMPLATE_PATH = Path(__file__).parent / "utils" / "model_card_template.md"
+MODEL_CARD_TEMPLATE_PATH = Path(__file__).parent / "model_card_template.md"
 SESSION_ID = uuid4().hex
 HF_HUB_OFFLINE = os.getenv("HF_HUB_OFFLINE", "").upper() in ENV_VARS_TRUE_VALUES
 DISABLE_TELEMETRY = os.getenv("DISABLE_TELEMETRY", "").upper() in ENV_VARS_TRUE_VALUES
@@ -87,10 +83,11 @@ def get_full_repo_name(model_id: str, organization: Optional[str] = None, token:
 
 
 def create_model_card(args, model_name):
-    if not is_modelcards_available:
+    if not is_jinja_available():
         raise ValueError(
-            "Please make sure to have `modelcards` installed when using the `create_model_card` function. You can"
-            " install the package with `pip install modelcards`."
+            "Modelcard rendering is based on Jinja templates."
+            " Please make sure to have `jinja` installed before using `create_model_card`."
+            " To install it, please run `pip install Jinja2`."
         )
 
     if hasattr(args, "local_rank") and args.local_rank not in [-1, 0]:
@@ -100,7 +97,7 @@ def create_model_card(args, model_name):
     repo_name = get_full_repo_name(model_name, token=hub_token)
 
     model_card = ModelCard.from_template(
-        card_data=CardData(  # Card metadata object that will be converted to YAML block
+        card_data=ModelCardData(  # Card metadata object that will be converted to YAML block
             language="en",
             license="apache-2.0",
             library_name="diffusers",
diff --git a/src/diffusers/utils/import_utils.py b/src/diffusers/utils/import_utils.py
@@ -21,6 +21,7 @@
 from collections import OrderedDict
 from typing import Union
 
+from huggingface_hub.utils import is_jinja_available  # noqa: F401
 from packaging import version
 from packaging.version import Version, parse
 
@@ -146,14 +147,6 @@
     _unidecode_available = False
 
 
-_modelcards_available = importlib.util.find_spec("modelcards") is not None
-try:
-    _modelcards_version = importlib_metadata.version("modelcards")
-    logger.debug(f"Successfully imported modelcards version {_modelcards_version}")
-except importlib_metadata.PackageNotFoundError:
-    _modelcards_available = False
-
-
 _onnxruntime_version = "N/A"
 _onnx_available = importlib.util.find_spec("onnxruntime") is not None
 if _onnx_available:
@@ -253,10 +246,6 @@ def is_unidecode_available():
     return _unidecode_available
 
 
-def is_modelcards_available():
-    return _modelcards_available
-
-
 def is_onnx_available():
     return _onnx_available
 
diff --git a/tests/test_hub_utils.py b/tests/test_hub_utils.py
@@ -0,0 +1,51 @@
+# coding=utf-8
+# Copyright 2023 HuggingFace Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+import unittest
+from pathlib import Path
+from tempfile import TemporaryDirectory
+from unittest.mock import Mock, patch
+
+import diffusers.utils.hub_utils
+
+
+class CreateModelCardTest(unittest.TestCase):
+    @patch("diffusers.utils.hub_utils.get_full_repo_name")
+    def test_create_model_card(self, repo_name_mock: Mock) -> None:
+        repo_name_mock.return_value = "full_repo_name"
+        with TemporaryDirectory() as tmpdir:
+            # Dummy args values
+            args = Mock()
+            args.output_dir = tmpdir
+            args.local_rank = 0
+            args.hub_token = "hub_token"
+            args.dataset_name = "dataset_name"
+            args.learning_rate = 0.01
+            args.train_batch_size = 100000
+            args.eval_batch_size = 10000
+            args.gradient_accumulation_steps = 0.01
+            args.adam_beta1 = 0.02
+            args.adam_beta2 = 0.03
+            args.adam_weight_decay = 0.0005
+            args.adam_epsilon = 0.000001
+            args.lr_scheduler = 1
+            args.lr_warmup_steps = 10
+            args.ema_inv_gamma = 0.001
+            args.ema_power = 0.1
+            args.ema_max_decay = 0.2
+            args.mixed_precision = True
+
+            # Model card mush be rendered and saved
+            diffusers.utils.hub_utils.create_model_card(args, model_name="model_name")
+            self.assertTrue((Path(tmpdir) / "README.md").is_file())
