# The timm Model API

Every `timm` model is a `torch.nn.Module` that shares a common interface on top of `forward()`. The interface splits
the model into a feature extractor and a classifier head, exposes intermediate features, and provides hooks for
training (gradient checkpointing, parameter grouping) and for changing the input resolution. Code written against it
works across architectures, which is how `timm`'s training script, feature extraction wrappers and optimizer factory
handle every model the same way.

This page describes the interface of a constructed model: which parts every model provides, which most provide, and
which are specific to some families. How models are created (`create_model` arguments, pretrained tags, the
pretrained config) is covered in the [models reference](reference/models).

## Feature extractor and head

A model's forward pass is `forward_head(forward_features(x))`.

- **`forward_features(x)`** returns the unpooled output of the last stage, after the final norm layer if the model
  has one: a `NCHW` feature map for convolutional models, `NLC` tokens (including any class / register prefix tokens)
  for ViT style models.
- **`forward_head(x, pre_logits=False)`** applies the global pooling (or token selection), any pre-logits layers and
  the classifier. With `pre_logits=True` it stops before the classifier and returns the pooled features.

Three attributes describe the sizes involved:

| Attribute | Description |
|---|---|
| `num_classes` | Classifier output size, `0` if the classifier was removed |
| `num_features` | Channels (or embedding dim) of the `forward_features` output |
| `head_hidden_size` | Size of the `pre_logits=True` output, the classifier input. Equal to `num_features` unless the head has a hidden layer (e.g. ConvNeXt w/ `head_hidden_size` set) |

```python
import torch
import timm

model = timm.create_model('convnext_atto').eval()
x = torch.randn(2, 3, 224, 224)

features = model.forward_features(x)  # (2, 320, 7, 7), num_features = 320
pooled = model.forward_head(features, pre_logits=True)  # (2, 320), head_hidden_size = 320
logits = model.forward_head(features)  # (2, 1000), same as model(x)
```

A few models deviate from the single tensor in, single tensor out form: multi-branch models (CrossViT,
EfficientViM) return a list from `forward_features`, and the NaFlex / Gemma4 ViTs accept extra patch coordinate and
validity inputs for variable resolution.

## Classifier

- **`get_classifier()`** returns the classifier module.
- **`reset_classifier(num_classes, global_pool=None)`** replaces the classifier with a new, randomly initialized one.
  `num_classes=0` replaces it with an identity, so `model(x)` returns the pooled features. `global_pool` changes the
  pooling at the same time, `''` disables it so `model(x)` returns unpooled features. In most models
  `global_pool=None` keeps the current pooling, but a few default to `'avg'`, pass it explicitly if it matters.

Creating a model with `num_classes` / `global_pool` set has the same effect as resetting the classifier afterwards,
without allocating the original classifier. Encoder-only vision towers (the DeepSeek, Gemma4, MobileNetV5 and Qwen3
encoders) have no classifier and don't implement these methods.

## Intermediate features

- **`forward_intermediates(x, indices=None, norm=False, stop_early=False, output_fmt='NCHW',
  intermediates_only=False)`** runs the model and returns the final features plus a list of intermediate features,
  or only the intermediates with `intermediates_only=True`. `indices` selects them: `None` for all, an `int` `n` for
  the last `n`, or a list of indices (negative values count from the end). `stop_early=True` skips the blocks after
  the last selected index (with `intermediates_only=True`), `norm=True` applies the final norm layer to each
  intermediate.
- **`feature_info`** describes the available intermediate features, one entry per index with the number of channels
  (`num_chs`), the stride relative to the input (`reduction`) and the module name (`module`).
- **`prune_intermediate_layers(indices=1, prune_norm=False, prune_head=True)`** removes the blocks after the last
  index that will be used (and optionally the final norm and the head) and returns the absolute indices to take.
  Use it when the model will only be used to extract intermediates, as a backbone for a detection or segmentation
  head.

```python
model = timm.create_model('resnet18').eval()
print([(f['num_chs'], f['reduction']) for f in model.feature_info])
# [(64, 2), (64, 4), (128, 8), (256, 16), (512, 32)]

final, intermediates = model.forward_intermediates(x)  # all 5 feature levels
last_two = model.forward_intermediates(x, indices=2, intermediates_only=True)  # (2, 256, 14, 14), (2, 512, 7, 7)

take_indices = model.prune_intermediate_layers(indices=2, prune_head=True)  # [3, 4], no classifier left
```

All features are returned as `NCHW` maps by default, including those of ViT style models, whose spatial tokens are
reshaped to a grid with the prefix tokens dropped. The other `output_fmt` options are family specific (`NLC` for
ViT style models, `NHWC` for some), as are extra arguments such as `return_prefix_tokens` (ViT family). Creating a
model with `features_only=True` wraps it as a backbone that returns the feature maps from `model(x)`, see
[feature extraction](feature_extraction) for both approaches in detail.

## Input size

Models fall into three groups by how they handle the input resolution:

- **Any input size.** Most convolutional models, and convolutional hybrids without fixed size position embeddings or
  window partitions, work at any input size larger than their total stride, without changes.
- **Fixed input size, resizable.** Models with position embeddings, attention window partitions or relative position
  tables are built for one size, set by the `img_size` argument at creation (default: the pretrained cfg
  `input_size`). In the families that implement it (ViT, DeiT, BEiT, EVA, Swin, MaxViT and others, see the table
  below), **`set_input_size(img_size=...)`** changes it afterwards, resampling the position embeddings and relative
  position tables and updating the windows. Several accept further arguments, e.g. `patch_size` for ViT,
  `window_size` / `window_ratio` for Swin. Calling it at the current size leaves the model unchanged. The others
  must be created with the `img_size` they will be used at.
- **Dynamic input size.** ViT and EVA family models created with `dynamic_img_size=True` resample their position
  embedding to the input size on every forward pass, at some cost in speed. Swin, SwinV2 and SwinV2-CR models created
  with `dynamic_img_size=True` (or the equivalent `strict_img_size=False`) compute their attention masks for the input
  size on each forward pass instead. NaFlex ViTs handle variable resolution natively.

```python
model = timm.create_model('vit_tiny_patch16_224')
model.set_input_size(img_size=(384, 384))
model(torch.randn(1, 3, 384, 384))
```

## Training hooks

- **`set_grad_checkpointing(enable=True)`** enables activation (gradient) checkpointing, recomputing the forward
  activations of each block in the backward pass to save memory (`--grad-checkpointing` in `train.py`). Most models
  support it, some older architectures (VGG, DPN, Inception, ...) raise an error.
- **`group_matcher(coarse=False)`** returns regular expressions that assign each parameter to a stem or block group.
  The optimizer factory uses it to scale learning rates per layer for layer-wise learning rate decay
  (`layer_decay`, see the [optimizers reference](reference/optimizers#parameter-groups)).
- **`no_weight_decay()`**, where implemented, returns the names of parameters that should not be weight decayed
  (position embeddings, class tokens, ...). The optimizer factory excludes these and all 1D parameters (biases,
  norm scales) from weight decay.

## Pretrained config

`model.pretrained_cfg` (and its older alias `model.default_cfg`) is a dict describing the source checkpoint: input
size, interpolation, crop, normalization, classifier and first conv module names, license and source. It is used to
adapt the classifier and input conv when pretrained weights are loaded, and as the preprocessing defaults. See the
[models reference](reference/models#pretrained-config) for the fields.

The pretrained cfg isn't changed when a model is created or modified to differ from its checkpoint (`in_chans`,
`img_size`, `set_input_size()`). The model's current input is tracked separately,
`timm.models.get_model_input_config(model)` returns its channels, size and fixed size constraint, and
`timm.data.resolve_input_data_config(model)` resolves a data config for the transforms against it.
`timm.models.get_model_args(model)` returns the arguments to rebuild the model as it is now, and `save_for_hf` writes
them to the exported config, so a resized or adapted model can be reloaded. See
[adapt and export models](model_adaptation).

## Support by family

The table below lists which optional parts of the interface each model family supports. It is generated by
`hfdocs/generate_model_docs.py`, which builds and runs a small model of each model class in each family. Support can
vary between the architectures of a family, see the family pages for details.

| Family | Classes | `features_only` | `forward_intermediates` | `prune_intermediate_layers` | Any input size | `set_input_size` | `dynamic_img_size` | Grad checkpointing | Layer decay | `reset_classifier` |
|---|---|---|---|---|---|---|---|---|---|---|
| [Aligned Xception](models/xception-aligned) | `XceptionAligned` | ✓ | ✗ | ✗ | ✓ | – | – | ✓ | ✓ | ✓ |
| [BEiT](models/beit) | `Beit` | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ |
| [BYOANet (BoTNet, HaloNet, LambdaNet)](models/byoanet) | `ByobNet` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [BYOBNet (GENet, RepVGG, MobileOne, ...)](models/byobnet) | `ByobNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [CaiT](models/cait) | `Cait` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [CoaT](models/coat) | `CoaT` | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ |
| [ConViT](models/convit) | `ConVit` | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ |
| [ConvMixer](models/convmixer) | `ConvMixer` | ✗ | ✗ | ✗ | ✓ | – | – | ✓ | ✓ | ✓ |
| [ConvNeXt](models/convnext) | `ConvNeXt` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [CPUBone](models/cpubone) | `CPUBone` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [CrossViT](models/crossvit) | `CrossVit` | ✗ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [CSATv2](models/csatv2) | `CSATv2` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✗ | ✓ |
| [CSPNet (CSPResNet, CSPDarkNet, DarkNet)](models/cspnet) | `CspNet` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [CSPNeXt](models/cspnext) | `CspNext` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [DaViT](models/davit) | `DaVit` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [DeepSeek ViT](models/deepseek-vit) | `DeepseekVitClassifier` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
|  | `DeepseekVitEncoder` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✗ |
| [DeiT](models/deit) | `VisionTransformer`, `VisionTransformerDistilled` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [DenseNet](models/densenet) | `DenseNet` | ✓ | ✗ | ✗ | ✓ | – | – | ✓ | ✓ | ✓ |
| [DLA](models/dla) | `DLA` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [DPN](models/dpn) | `DPN` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [EdgeNeXt](models/edgenext) | `EdgeNeXt` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [EfficientFormer](models/efficientformer) | `EfficientFormer` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [EfficientFormer-V2](models/efficientformer-v2) | `EfficientFormerV2` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [EfficientNet](models/efficientnet) | `EfficientNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [EfficientViM](models/efficientvim) | `EfficientViM`, `EfficientViMDistilled` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [EfficientViT (MIT)](models/efficientvit-mit) | `EfficientVit`, `EfficientVitLarge` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [EfficientViT (MSRA)](models/efficientvit-msra) | `EfficientVitMsra` | ✓ | ✓ | ✓ | ✓ | ✓ | – | ✓ | ✓ | ✓ |
| [EVA](models/eva) | `Eva` | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [FasterNet](models/fasternet) | `FasterNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [FastViT](models/fastvit) | `FastVit` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [FocalNet](models/focalnet) | `FocalNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [GCViT](models/gcvit) | `GlobalContextVit` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [Gemma 4 ViT](models/gemma4-vit) | `Gemma4VitClassifier` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
|  | `Gemma4VitEncoder` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✗ |
| [GhostNet](models/ghostnet) | `GhostNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [HardCoRe-NAS](models/hardcorenas) | `MobileNetV3` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Hiera](models/hiera) | `Hiera` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [HieraDet (SAM 2)](models/hieradet-sam2) | `HieraDet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [HRNet](models/hrnet) | `HighResolutionNet` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [Hybrid Vision Transformer](models/vision-transformer-hybrid) | `VisionTransformer` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [iFormer](models/iformer) | `IFormer` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Inception-ResNet-v2](models/inception-resnet-v2) | `InceptionResnetV2` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [Inception-v3](models/inception-v3) | `InceptionV3` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [Inception-v4](models/inception-v4) | `InceptionV4` | ✓ | ✓ | ✓ | ✓ | – | – | ✗ | ✓ | ✓ |
| [InceptionNeXt](models/inception-next) | `MetaNeXt` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [LeViT](models/levit) | `Levit`, `LevitDistilled` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [LowFormer](models/lowformer) | `LowFormer` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [MambaOut](models/mambaout) | `MambaOut` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [MaxViT and CoAtNet](models/maxxvit) | `MaxxVit` | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ |
| [MetaFormer (PoolFormer, ConvFormer, CAFormer)](models/metaformer) | `MetaFormer` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✗ | ✓ |
| [MLP-Mixer, ResMLP and gMLP](models/mlp-mixer) | `MlpMixer` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [MobileNet-V3](models/mobilenetv3) | `MobileNetV3` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [MobileNet-V5](models/mobilenetv5) | `MobileNetV5` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
|  | `MobileNetV5Encoder` | ✓ | ✓ | ✗ | ✓ | – | – | ✓ | ✗ | ✗ |
| [MobileViT](models/mobilevit) | `ByobNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [MViTv2](models/mvitv2) | `MultiScaleVit` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [NaFlex ViT](models/naflexvit) | `NaFlexVit` | ✓ | ✓ | ✗ | ✓ | – | – | ✓ | ✓ | ✓ |
| [NASNet](models/nasnet) | `NASNetALarge` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [NesT](models/nest) | `Nest` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [Next-ViT](models/nextvit) | `NextViT` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [NFNet](models/nfnet) | `NormFreeNet` | ✓ | ✗ | ✗ | ✓ | – | – | ✓ | ✓ | ✓ |
| [PiT](models/pit) | `PoolingVisionTransformer` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ |
| [PNASNet](models/pnasnet) | `PNASNet5Large` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [PP-HGNet](models/hgnet) | `HighPerfGpuNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [PP-LCNetV2](models/lcnetv2) | `LCNetV2` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [PVTv2](models/pvt-v2) | `PyramidVisionTransformerV2` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Qwen3 ViT](models/qwen3-vit) | `Qwen3VitClassifier` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
|  | `Qwen3VitEncoder` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✗ |
| [RDNet](models/rdnet) | `RDNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [RegNet](models/regnet) | `RegNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [RepGhost](models/repghost) | `RepGhostNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [RepViT](models/repvit) | `RepVit` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Res2Net](models/res2net) | `ResNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [ResNeSt](models/resnest) | `ResNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [ResNet](models/resnet) | `ResNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [ResNetV2 (BiT)](models/resnetv2) | `ResNetV2` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [ReXNet](models/rexnet) | `RexNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Segment Anything ViT](models/vision-transformer-sam) | `VisionTransformerSAM` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [SelecSLS](models/selecsls) | `SelecSls` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [SENet (legacy)](models/senet) | `SENet` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [Sequencer](models/sequencer) | `Sequencer2d` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [SHViT](models/shvit) | `SHViT` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [SKNet](models/sknet) | `ResNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [StarNet](models/starnet) | `StarNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [SwiftFormer](models/swiftformer) | `SwiftFormer` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Swin Transformer](models/swin-transformer) | `SwinTransformer` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [Swin Transformer V2](models/swin-transformer-v2) | `SwinTransformerV2` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [Swin Transformer V2 (CR)](models/swin-transformer-v2-cr) | `SwinTransformerV2Cr` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [TinyViT](models/tiny-vit) | `TinyVit` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [TNT](models/tnt) | `TNT` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [TResNet](models/tresnet) | `TResNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Twins](models/twins) | `Twins` | ✓ | ✓ | ✓ | ✓ | – | – | ✗ | ✓ | ✓ |
| [VGG](models/vgg) | `VGG` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [Visformer](models/visformer) | `Visformer` | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [Vision Transformer (relative position)](models/vision-transformer-relpos) | `VisionTransformerRelPos` | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ | ✓ | ✓ | ✓ |
| [Vision Transformer (ViT)](models/vision-transformer) | `VisionTransformer` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [ViTamin](models/vitamin) | `VisionTransformer` | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ |
| [VOLO](models/volo) | `VOLO` | ✓ | ✓ | ✓ | ✗ | ✗ | ✗ | ✓ | ✓ | ✓ |
| [VoVNet](models/vovnet) | `VovNet` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |
| [Xception (legacy)](models/xception) | `Xception` | ✓ | ✗ | ✗ | ✓ | – | – | ✗ | ✓ | ✓ |
| [XCiT](models/xcit) | `Xcit` | ✓ | ✓ | ✓ | ✓ | – | – | ✓ | ✓ | ✓ |

✓ supported, ✓ (square) works for square sizes only, ✗ not supported, – not needed (the model already accepts any input
size). Each model class of a family is checked by running every feature on a randomly initialized model of its smallest
(non-test) architecture in the family. Classes with the same results share a row, the family page lists the architecture
used for each.

## Implementing a model

A new model should implement the interface so it works with the training script, feature extraction and the
optimizer factory without special cases:

- `forward_features`, `forward_head(x, pre_logits=False)` and `forward` as their composition, with `num_classes`,
  `num_features` and `head_hidden_size` set.
- `get_classifier` and `reset_classifier(num_classes, global_pool=None)`.
- `forward_intermediates`, `prune_intermediate_layers` and a `feature_info` list.
- `set_grad_checkpointing`, `group_matcher`, and `no_weight_decay` for any parameter that shouldn't be decayed.
- `set_input_size` if the model is built for a fixed input size.
- The standard constructor arguments: `num_classes`, `in_chans`, `global_pool`, `drop_rate`, `drop_path_rate`,
  and `device` / `dtype` for parameter creation.

`tests/test_models.py` checks the forward, backward, features, intermediates, classifier reset and pretrained cfg
parts of the interface for the registered models (the largest are excluded from some tests to keep CI time down).