Merge remote-tracking branch 'origin/main' into multi-weight

.github/workflows/tests.yml

@@ -15,10 +15,10 @@ jobs:
     name: Run tests on ${{ matrix.os }} with Python ${{ matrix.python }}
     strategy:
       matrix:
-        os: [ubuntu-latest, macOS-latest]
+        os: [ubuntu-latest]
-        python: ['3.9']
+        python: ['3.10']
-        torch: ['1.10.0']
+        torch: ['1.13.0']
-        torchvision: ['0.11.1']
+        torchvision: ['0.14.0']
     runs-on: ${{ matrix.os }}
 
     steps:
@@ -34,6 +34,9 @@ jobs:
     - name: Install torch on mac
       if: startsWith(matrix.os, 'macOS')
       run: pip install --no-cache-dir torch==${{ matrix.torch }} torchvision==${{ matrix.torchvision }}
+    - name: Install torch on Windows
+      if: startsWith(matrix.os, 'windows')
+      run: pip install --no-cache-dir torch==${{ matrix.torch }} torchvision==${{ matrix.torchvision }}
     - name: Install torch on ubuntu
       if: startsWith(matrix.os, 'ubuntu')
       run: |
@@ -42,11 +45,18 @@ jobs:
         sudo apt install -y google-perftools
     - name: Install requirements
       run: |
-        if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
+        pip install -r requirements.txt
-        pip install --no-cache-dir git+https://github.com/mapillary/inplace_abn.git@v1.1.0
+        pip install --no-cache-dir git+https://github.com/mapillary/inplace_abn.git
-    - name: Run tests
+    - name: Run tests on Windows
+      if: startsWith(matrix.os, 'windows')
+      env:
+        PYTHONDONTWRITEBYTECODE: 1
+      run: |
+        pytest -vv tests
+    - name: Run tests on Linux / Mac
+      if: ${{ !startsWith(matrix.os, 'windows') }}
       env:
         LD_PRELOAD: /usr/lib/x86_64-linux-gnu/libtcmalloc.so.4
+        PYTHONDONTWRITEBYTECODE: 1
       run: |
-        export PYTHONDONTWRITEBYTECODE=1
+        pytest -vv --forked --durations=0 tests
-        pytest -vv --forked --durations=0 ./tests

timm/data/dataset_factory.py

@@ -4,7 +4,7 @@ Hacked together by / Copyright 2021, Ross Wightman
 """
 import os
 
-from torchvision.datasets import CIFAR100, CIFAR10, MNIST, QMNIST, KMNIST, FashionMNIST, ImageNet, ImageFolder
+from torchvision.datasets import CIFAR100, CIFAR10, MNIST, KMNIST, FashionMNIST, ImageFolder
 try:
     from torchvision.datasets import Places365
     has_places365 = True
@@ -15,6 +15,16 @@ try:
     has_inaturalist = True
 except ImportError:
     has_inaturalist = False
+try:
+    from torchvision.datasets import QMNIST
+    has_qmnist = True
+except ImportError:
+    has_qmnist = False
+try:
+    from torchvision.datasets import ImageNet
+    has_imagenet = True
+except ImportError:
+    has_imagenet = False
 
 from .dataset import IterableImageDataset, ImageDataset
 
@@ -22,7 +32,6 @@ _TORCH_BASIC_DS = dict(
     cifar10=CIFAR10,
     cifar100=CIFAR100,
     mnist=MNIST,
-    qmist=QMNIST,
     kmnist=KMNIST,
     fashion_mnist=FashionMNIST,
 )
@@ -122,7 +131,12 @@ def create_dataset(
             elif split in _EVAL_SYNONYM:
                 split = 'val'
             ds = Places365(split=split, **torch_kwargs)
+        elif name == 'qmnist':
+            assert has_qmnist, 'Please update to a newer PyTorch and torchvision for QMNIST dataset.'
+            use_train = split in _TRAIN_SYNONYM
+            ds = QMNIST(train=use_train, **torch_kwargs)
         elif name == 'imagenet':
+            assert has_imagenet, 'Please update to a newer PyTorch and torchvision for ImageNet dataset.'
             if split in _EVAL_SYNONYM:
                 split = 'val'
             ds = ImageNet(split=split, **torch_kwargs)

timm/models/hub.py

@@ -3,10 +3,12 @@ import logging
 import os
 from functools import partial
 from pathlib import Path
-from typing import Union
+from tempfile import TemporaryDirectory
+from typing import Optional, Union
 
 import torch
 from torch.hub import HASH_REGEX, download_url_to_file, urlparse
+
 try:
     from torch.hub import get_dir
 except ImportError:
@@ -15,7 +17,10 @@ except ImportError:
 from timm import __version__
 
 try:
-    from huggingface_hub import HfApi, HfFolder, Repository, hf_hub_download, hf_hub_url
+    from huggingface_hub import (create_repo, get_hf_file_metadata,
+                                 hf_hub_download, hf_hub_url,
+                                 repo_type_and_id_from_hf_id, upload_folder)
+    from huggingface_hub.utils import EntryNotFoundError
     hf_hub_download = partial(hf_hub_download, library_name="timm", library_version=__version__)
     _has_hf_hub = True
 except ImportError:
@@ -121,53 +126,45 @@ def save_for_hf(model, save_directory, model_config=None):
 
 def push_to_hf_hub(
     model,
-    local_dir,
+    repo_id: str,
-    repo_namespace_or_url=None,
+    commit_message: str ='Add model',
-    commit_message='Add model',
+    token: Optional[str] = None,
-    use_auth_token=True,
+    revision: Optional[str] = None,
-    git_email=None,
+    private: bool = False,
-    git_user=None,
+    create_pr: bool = False,
-    revision=None,
+    model_config: Optional[dict] = None,
-    model_config=None,
 ):
-    if repo_namespace_or_url:
+    # Create repo if doesn't exist yet
-        repo_owner, repo_name = repo_namespace_or_url.rstrip('/').split('/')[-2:]
+    repo_url = create_repo(repo_id, token=token, private=private, exist_ok=True)
-    else:
+
-        if isinstance(use_auth_token, str):
+    # Infer complete repo_id from repo_url
-            token = use_auth_token
+    # Can be different from the input `repo_id` if repo_owner was implicit
-        else:
+    _, repo_owner, repo_name = repo_type_and_id_from_hf_id(repo_url)
-            token = HfFolder.get_token()
+    repo_id = f"{repo_owner}/{repo_name}"
-
+
-        if token is None:
+    # Check if README file already exist in repo
-            raise ValueError(
+    try:
-                "You must login to the Hugging Face hub on this computer by typing `transformers-cli login` and "
+        get_hf_file_metadata(hf_hub_url(repo_id=repo_id, filename="README.md", revision=revision))
-                "entering your credentials to use `use_auth_token=True`. Alternatively, you can pass your own "
+        has_readme = True
-                "token as the `use_auth_token` argument."
+    except EntryNotFoundError:
-            )
+        has_readme = False
-
+
-        repo_owner = HfApi().whoami(token)['name']
+    # Dump model and push to Hub
-        repo_name = Path(local_dir).name
+    with TemporaryDirectory() as tmpdir:
-
-    repo_url = f'https://huggingface.co/{repo_owner}/{repo_name}'
-
-    repo = Repository(
-        local_dir,
-        clone_from=repo_url,
-        use_auth_token=use_auth_token,
-        git_user=git_user,
-        git_email=git_email,
-        revision=revision,
-    )
-
-    # Prepare a default model card that includes the necessary tags to enable inference.
-    readme_text = f'---\ntags:\n- image-classification\n- timm\nlibrary_tag: timm\n---\n# Model card for {repo_name}'
-    with repo.commit(commit_message):
         # Save model weights and config.
-        save_for_hf(model, repo.local_dir, model_config=model_config)
+        save_for_hf(model, tmpdir, model_config=model_config)
 
-        # Save a model card if it doesn't exist.
+        # Add readme if does not exist
-        readme_path = Path(repo.local_dir) / 'README.md'
+        if not has_readme:
-        if not readme_path.exists():
+            readme_path = Path(tmpdir) / "README.md"
+            readme_text = f'---\ntags:\n- image-classification\n- timm\nlibrary_tag: timm\n---\n# Model card for {repo_id}'
             readme_path.write_text(readme_text)
 
-    return repo.git_remote_url()
+        # Upload model and return
+        return upload_folder(
+            repo_id=repo_id,
+            folder_path=tmpdir,
+            revision=revision,
+            create_pr=create_pr,
+            commit_message=commit_message,
+        )

timm/models/layers/norm.py

@@ -50,7 +50,7 @@ class LayerNorm(nn.LayerNorm):
 
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         if self._fast_norm:
-           x = fast_layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
+            x = fast_layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
         else:
             x = F.layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
         return x
@@ -65,7 +65,7 @@ class LayerNorm2d(nn.LayerNorm):
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         x = x.permute(0, 2, 3, 1)
         if self._fast_norm:
-           x = fast_layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
+            x = fast_layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
         else:
             x = F.layer_norm(x, self.normalized_shape, self.weight, self.bias, self.eps)
         x = x.permute(0, 3, 1, 2)

train.py

@@ -104,7 +104,7 @@ group.add_argument('--class-map', default='', type=str, metavar='FILENAME',
 # Model parameters
 group = parser.add_argument_group('Model parameters')
 group.add_argument('--model', default='resnet50', type=str, metavar='MODEL',
-                    help='Name of model to train (default: "resnet50"')
+                    help='Name of model to train (default: "resnet50")')
 group.add_argument('--pretrained', action='store_true', default=False,
                     help='Start with pretrained version of specified network (if avail)')
 group.add_argument('--initial-checkpoint', default='', type=str, metavar='PATH',
@@ -157,7 +157,7 @@ scripting_group.add_argument('--dynamo', default=False, action='store_true',
 # Optimizer parameters
 group = parser.add_argument_group('Optimizer parameters')
 group.add_argument('--opt', default='sgd', type=str, metavar='OPTIMIZER',
-                    help='Optimizer (default: "sgd"')
+                    help='Optimizer (default: "sgd")')
 group.add_argument('--opt-eps', default=None, type=float, metavar='EPSILON',
                     help='Optimizer Epsilon (default: None, use opt default)')
 group.add_argument('--opt-betas', default=None, type=float, nargs='+', metavar='BETA',
@@ -222,7 +222,7 @@ group.add_argument('--warmup-prefix', action='store_true', default=False,
 group.add_argument('--cooldown-epochs', type=int, default=0, metavar='N',
                     help='epochs to cooldown LR at min_lr, after cyclic schedule ends')
 group.add_argument('--patience-epochs', type=int, default=10, metavar='N',
-                    help='patience epochs for Plateau LR scheduler (default: 10')
+                    help='patience epochs for Plateau LR scheduler (default: 10)')
 group.add_argument('--decay-rate', '--dr', type=float, default=0.1, metavar='RATE',
                     help='LR decay rate (default: 0.1)')
 
@@ -419,7 +419,7 @@ def main():
 
     in_chans = 3
     if args.in_chans is not None:
-        in_chans = args.in_chanes
+        in_chans = args.in_chans
     elif args.input_size is not None:
         in_chans = args.input_size[0]