add MobileNetV1 model (#17799)

* add model files etc for MobileNetV2

rename files for MobileNetV1

initial implementation of MobileNetV1

fix conversion script

cleanup

write docs

tweaks

fix conversion script

extract hidden states

fix test cases

make fixup

fixup it all

remove main from doc link

fixes

fix tests

fix up

use google org

fix weird assert

* fixup

* use google organization for checkpoints

README.md

@@ -344,6 +344,7 @@ Current number of checkpoints: ![](https://img.shields.io/endpoint?url=https://h
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.

README_es.md

@@ -344,6 +344,7 @@ Número actual de puntos de control: ![](https://img.shields.io/endpoint?url=htt
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.

README_ja.md

@@ -379,6 +379,7 @@ Flax、PyTorch、TensorFlowをcondaでインストールする方法は、それ
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.

README_ko.md

@@ -294,6 +294,7 @@ Flax, PyTorch, TensorFlow 설치 페이지에서 이들을 conda로 설치하는
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.

README_zh-hans.md

@@ -318,6 +318,7 @@ conda install -c huggingface transformers
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (来自 NVIDIA) 伴随论文 [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) 由 Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro 发布。
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (来自 Studio Ousia) 伴随论文 [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) 由 Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka 发布。
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (来自 CMU/Google Brain) 伴随论文 [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) 由 Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou 发布。
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (来自 Google Inc.) 伴随论文 [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) 由 Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam 发布。
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (来自 Google Inc.) 伴随论文 [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) 由 Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen 发布。
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (来自 Apple) 伴随论文 [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) 由 Sachin Mehta and Mohammad Rastegari 发布。
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (来自 Microsoft Research) 伴随论文 [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) 由 Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu 发布。

README_zh-hant.md

@@ -330,6 +330,7 @@ conda install -c huggingface transformers
 1. **[Megatron-GPT2](https://huggingface.co/docs/transformers/model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](https://huggingface.co/docs/transformers/model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](https://huggingface.co/docs/transformers/model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](https://huggingface.co/docs/transformers/model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](https://huggingface.co/docs/transformers/model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](https://huggingface.co/docs/transformers/model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](https://huggingface.co/docs/transformers/model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.

docs/source/en/_toctree.yml

@@ -410,6 +410,8 @@
         title: LeViT
       - local: model_doc/maskformer
         title: MaskFormer
+      - local: model_doc/mobilenet_v1
+        title: MobileNetV1
       - local: model_doc/mobilenet_v2
         title: MobileNetV2
       - local: model_doc/mobilevit

docs/source/en/index.mdx

@@ -132,6 +132,7 @@ The documentation is organized into five sections:
 1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
 1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
 1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
+1. **[MobileNetV1](model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
 1. **[MobileNetV2](model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
 1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
 1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
@@ -288,6 +289,7 @@ Flax), PyTorch, and/or TensorFlow.
 |            mBART            |       ✅       |       ✅       |       ✅        |         ✅         |      ✅      |
 |        Megatron-BERT        |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         MobileBERT          |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |
+|         MobileNetV1         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |         MobileNetV2         |       ❌       |       ❌       |       ✅        |         ❌         |      ❌      |
 |          MobileViT          |       ❌       |       ❌       |       ✅        |         ✅         |      ❌      |
 |            MPNet            |       ✅       |       ✅       |       ✅        |         ✅         |      ❌      |

docs/source/en/model_doc/mobilenet_v1.mdx

+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+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.
+-->
+
+# MobileNet V1
+
+## Overview
+
+The MobileNet model was proposed in [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
+
+The abstract from the paper is the following:
+
+*We present a class of efficient models called MobileNets for mobile and embedded vision applications. MobileNets are based on a streamlined architecture that uses depth-wise separable convolutions to build light weight deep neural networks. We introduce two simple global hyper-parameters that efficiently trade off between latency and accuracy. These hyper-parameters allow the model builder to choose the right sized model for their application based on the constraints of the problem. We present extensive experiments on resource and accuracy tradeoffs and show strong performance compared to other popular models on ImageNet classification. We then demonstrate the effectiveness of MobileNets across a wide range of applications and use cases including object detection, finegrain classification, face attributes and large scale geo-localization.*
+
+Tips:
+
+- The checkpoints are named **mobilenet\_v1\_*depth*\_*size***, for example **mobilenet\_v1\_1.0\_224**, where **1.0** is the depth multiplier (sometimes also referred to as "alpha" or the width multiplier) and **224** is the resolution of the input images the model was trained on.
+
+- Even though the checkpoint is trained on images of specific size, the model will work on images of any size. The smallest supported image size is 32x32.
+
+- One can use [`MobileNetV1FeatureExtractor`] to prepare images for the model.
+
+- The available image classification checkpoints are pre-trained on [ImageNet-1k](https://huggingface.co/datasets/imagenet-1k) (also referred to as ILSVRC 2012, a collection of 1.3 million images and 1,000 classes). However, the model predicts 1001 classes: the 1000 classes from ImageNet plus an extra “background” class (index 0).
+
+- The original TensorFlow checkpoints use different padding rules than PyTorch, requiring the model to determine the padding amount at inference time, since this depends on the input image size. To use native PyTorch padding behavior, create a [`MobileNetV1Config`] with `tf_padding = False`.
+
+Unsupported features:
+
+- The [`MobileNetV1Model`] outputs a globally pooled version of the last hidden state. In the original model it is possible to use a 7x7 average pooling layer with stride 2 instead of global pooling. For larger inputs, this gives a pooled output that is larger than 1x1 pixel. The HuggingFace implementation does not support this.
+
+- It is currently not possible to specify an `output_stride`. For smaller output strides, the original model invokes dilated convolution to prevent the spatial resolution from being reduced further. The output stride of the HuggingFace model is always 32.
+
+- The original TensorFlow checkpoints include quantized models. We do not support these models as they include additional "FakeQuantization" operations to unquantize the weights.
+
+- It's common to extract the output from the pointwise layers at indices 5, 11, 12, 13 for downstream purposes. Using `output_hidden_states=True` returns the output from all intermediate layers. There is currently no way to limit this to specific layers.
+
+This model was contributed by [matthijs](https://huggingface.co/Matthijs). The original code and weights can be found [here](https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet_v1.md).
+
+## MobileNetV1Config
+
+[[autodoc]] MobileNetV1Config
+
+## MobileNetV1FeatureExtractor
+
+[[autodoc]] MobileNetV1FeatureExtractor
+    - preprocess
+
+## MobileNetV1ImageProcessor
+
+[[autodoc]] MobileNetV1ImageProcessor
+    - preprocess
+
+## MobileNetV1Model
+
+[[autodoc]] MobileNetV1Model
+    - forward
+
+## MobileNetV1ForImageClassification
+
+[[autodoc]] MobileNetV1ForImageClassification
+    - forward

docs/source/en/serialization.mdx

@@ -84,6 +84,7 @@ Ready-made configurations include the following architectures:
 - Marian
 - mBART
 - MobileBERT
+- MobileNetV1
 - MobileNetV2
 - MobileViT
 - MT5

src/transformers/__init__.py

@@ -300,6 +300,7 @@ _import_structure = {
     "models.mluke": [],
     "models.mmbt": ["MMBTConfig"],
     "models.mobilebert": ["MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP", "MobileBertConfig", "MobileBertTokenizer"],
+    "models.mobilenet_v1": ["MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP", "MobileNetV1Config"],
     "models.mobilenet_v2": ["MOBILENET_V2_PRETRAINED_CONFIG_ARCHIVE_MAP", "MobileNetV2Config"],
     "models.mobilevit": ["MOBILEVIT_PRETRAINED_CONFIG_ARCHIVE_MAP", "MobileViTConfig"],
     "models.mpnet": ["MPNET_PRETRAINED_CONFIG_ARCHIVE_MAP", "MPNetConfig", "MPNetTokenizer"],
@@ -730,6 +731,7 @@ else:
     _import_structure["models.layoutlmv3"].extend(["LayoutLMv3FeatureExtractor", "LayoutLMv3ImageProcessor"])
     _import_structure["models.levit"].extend(["LevitFeatureExtractor", "LevitImageProcessor"])
     _import_structure["models.maskformer"].append("MaskFormerFeatureExtractor")
+    _import_structure["models.mobilenet_v1"].extend(["MobileNetV1FeatureExtractor", "MobileNetV1ImageProcessor"])
     _import_structure["models.mobilenet_v2"].extend(["MobileNetV2FeatureExtractor", "MobileNetV2ImageProcessor"])
     _import_structure["models.mobilevit"].extend(["MobileViTFeatureExtractor", "MobileViTImageProcessor"])
     _import_structure["models.owlvit"].append("OwlViTFeatureExtractor")
@@ -1668,6 +1670,15 @@ else:
             "load_tf_weights_in_mobilebert",
         ]
     )
+    _import_structure["models.mobilenet_v1"].extend(
+        [
+            "MOBILENET_V1_PRETRAINED_MODEL_ARCHIVE_LIST",
+            "MobileNetV1ForImageClassification",
+            "MobileNetV1Model",
+            "MobileNetV1PreTrainedModel",
+            "load_tf_weights_in_mobilenet_v1",
+        ]
+    )
     _import_structure["models.mobilenet_v2"].extend(
         [
             "MOBILENET_V2_PRETRAINED_MODEL_ARCHIVE_LIST",
@@ -3456,6 +3467,7 @@ if TYPE_CHECKING:
     from .models.megatron_bert import MEGATRON_BERT_PRETRAINED_CONFIG_ARCHIVE_MAP, MegatronBertConfig
     from .models.mmbt import MMBTConfig
     from .models.mobilebert import MOBILEBERT_PRETRAINED_CONFIG_ARCHIVE_MAP, MobileBertConfig, MobileBertTokenizer
+    from .models.mobilenet_v1 import MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP, MobileNetV1Config
     from .models.mobilenet_v2 import MOBILENET_V2_PRETRAINED_CONFIG_ARCHIVE_MAP, MobileNetV2Config
     from .models.mobilevit import MOBILEVIT_PRETRAINED_CONFIG_ARCHIVE_MAP, MobileViTConfig
     from .models.mpnet import MPNET_PRETRAINED_CONFIG_ARCHIVE_MAP, MPNetConfig, MPNetTokenizer
@@ -3830,6 +3842,7 @@ if TYPE_CHECKING:
         from .models.layoutlmv3 import LayoutLMv3FeatureExtractor, LayoutLMv3ImageProcessor
         from .models.levit import LevitFeatureExtractor, LevitImageProcessor
         from .models.maskformer import MaskFormerFeatureExtractor
+        from .models.mobilenet_v1 import MobileNetV1FeatureExtractor, MobileNetV1ImageProcessor
         from .models.mobilenet_v2 import MobileNetV2FeatureExtractor, MobileNetV2ImageProcessor
         from .models.mobilevit import MobileViTFeatureExtractor, MobileViTImageProcessor
         from .models.owlvit import OwlViTFeatureExtractor
@@ -4572,6 +4585,13 @@ if TYPE_CHECKING:
             MobileBertPreTrainedModel,
             load_tf_weights_in_mobilebert,
         )
+        from .models.mobilenet_v1 import (
+            MOBILENET_V1_PRETRAINED_MODEL_ARCHIVE_LIST,
+            MobileNetV1ForImageClassification,
+            MobileNetV1Model,
+            MobileNetV1PreTrainedModel,
+            load_tf_weights_in_mobilenet_v1,
+        )
         from .models.mobilenet_v2 import (
             MOBILENET_V2_PRETRAINED_MODEL_ARCHIVE_LIST,
             MobileNetV2ForImageClassification,

src/transformers/models/__init__.py

@@ -103,6 +103,7 @@ from . import (
     mluke,
     mmbt,
     mobilebert,
+    mobilenet_v1,
     mobilenet_v2,
     mobilevit,
     mpnet,

src/transformers/models/auto/configuration_auto.py

@@ -101,6 +101,7 @@ CONFIG_MAPPING_NAMES = OrderedDict(
         ("mctct", "MCTCTConfig"),
         ("megatron-bert", "MegatronBertConfig"),
         ("mobilebert", "MobileBertConfig"),
+        ("mobilenet_v1", "MobileNetV1Config"),
         ("mobilenet_v2", "MobileNetV2Config"),
         ("mobilevit", "MobileViTConfig"),
         ("mpnet", "MPNetConfig"),
@@ -244,6 +245,7 @@ CONFIG_ARCHIVE_MAP_MAPPING_NAMES = OrderedDict(
         ("mbart", "MBART_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("mctct", "MCTCT_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("megatron-bert", "MEGATRON_BERT_PRETRAINED_CONFIG_ARCHIVE_MAP"),
+        ("mobilenet_v1", "MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("mobilenet_v2", "MOBILENET_V2_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("mobilevit", "MOBILEVIT_PRETRAINED_CONFIG_ARCHIVE_MAP"),
         ("mpnet", "MPNET_PRETRAINED_CONFIG_ARCHIVE_MAP"),
@@ -397,6 +399,7 @@ MODEL_NAMES_MAPPING = OrderedDict(
         ("megatron_gpt2", "Megatron-GPT2"),
         ("mluke", "mLUKE"),
         ("mobilebert", "MobileBERT"),
+        ("mobilenet_v1", "MobileNetV1"),
         ("mobilenet_v2", "MobileNetV2"),
         ("mobilevit", "MobileViT"),
         ("mpnet", "MPNet"),

src/transformers/models/auto/feature_extraction_auto.py

@@ -61,6 +61,7 @@ FEATURE_EXTRACTOR_MAPPING_NAMES = OrderedDict(
         ("levit", "LevitFeatureExtractor"),
         ("maskformer", "MaskFormerFeatureExtractor"),
         ("mctct", "MCTCTFeatureExtractor"),
+        ("mobilenet_v1", "MobileNetV1FeatureExtractor"),
         ("mobilenet_v2", "MobileNetV2FeatureExtractor"),
         ("mobilevit", "MobileViTFeatureExtractor"),
         ("nat", "ViTFeatureExtractor"),

src/transformers/models/auto/image_processing_auto.py

@@ -52,6 +52,7 @@ IMAGE_PROCESSOR_MAPPING_NAMES = OrderedDict(
         ("layoutlmv2", "LayoutLMv2ImageProcessor"),
         ("layoutlmv3", "LayoutLMv3ImageProcessor"),
         ("levit", "LevitImageProcessor"),
+        ("mobilenet_v1", "MobileNetV1ImageProcessor"),
         ("mobilenet_v2", "MobileNetV2ImageProcessor"),
         ("mobilevit", "MobileViTImageProcessor"),
         ("nat", "ViTImageProcessor"),

src/transformers/models/auto/modeling_auto.py

@@ -100,6 +100,7 @@ MODEL_MAPPING_NAMES = OrderedDict(
         ("mctct", "MCTCTModel"),
         ("megatron-bert", "MegatronBertModel"),
         ("mobilebert", "MobileBertModel"),
+        ("mobilenet_v1", "MobileNetV1Model"),
         ("mobilenet_v2", "MobileNetV2Model"),
         ("mobilevit", "MobileViTModel"),
         ("mpnet", "MPNetModel"),
@@ -372,6 +373,7 @@ MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES = OrderedDict(
         ("dinat", "DinatForImageClassification"),
         ("imagegpt", "ImageGPTForImageClassification"),
         ("levit", ("LevitForImageClassification", "LevitForImageClassificationWithTeacher")),
+        ("mobilenet_v1", "MobileNetV1ForImageClassification"),
         ("mobilenet_v2", "MobileNetV2ForImageClassification"),
         ("mobilevit", "MobileViTForImageClassification"),
         ("nat", "NatForImageClassification"),

src/transformers/models/mobilenet_v1/__init__.py

+# flake8: noqa
+# There's no way to ignore "F401 '...' imported but unused" warnings in this
+# module, but to preserve other warnings. So, don't check this module at all.
+
+# Copyright 2022 The HuggingFace Team. All rights reserved.
+#
+# 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.
+from typing import TYPE_CHECKING
+
+from ...utils import OptionalDependencyNotAvailable, _LazyModule, is_torch_available, is_vision_available
+
+
+_import_structure = {
+    "configuration_mobilenet_v1": [
+        "MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP",
+        "MobileNetV1Config",
+        "MobileNetV1OnnxConfig",
+    ],
+}
+
+try:
+    if not is_vision_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["feature_extraction_mobilenet_v1"] = ["MobileNetV1FeatureExtractor"]
+    _import_structure["image_processing_mobilenet_v1"] = ["MobileNetV1ImageProcessor"]
+
+try:
+    if not is_torch_available():
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    pass
+else:
+    _import_structure["modeling_mobilenet_v1"] = [
+        "MOBILENET_V1_PRETRAINED_MODEL_ARCHIVE_LIST",
+        "MobileNetV1ForImageClassification",
+        "MobileNetV1Model",
+        "MobileNetV1PreTrainedModel",
+        "load_tf_weights_in_mobilenet_v1",
+    ]
+
+
+if TYPE_CHECKING:
+    from .configuration_mobilenet_v1 import (
+        MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP,
+        MobileNetV1Config,
+        MobileNetV1OnnxConfig,
+    )
+
+    try:
+        if not is_vision_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .feature_extraction_mobilenet_v1 import MobileNetV1FeatureExtractor
+        from .image_processing_mobilenet_v1 import MobileNetV1ImageProcessor
+
+    try:
+        if not is_torch_available():
+            raise OptionalDependencyNotAvailable()
+    except OptionalDependencyNotAvailable:
+        pass
+    else:
+        from .modeling_mobilenet_v1 import (
+            MOBILENET_V1_PRETRAINED_MODEL_ARCHIVE_LIST,
+            MobileNetV1ForImageClassification,
+            MobileNetV1Model,
+            MobileNetV1PreTrainedModel,
+            load_tf_weights_in_mobilenet_v1,
+        )
+
+
+else:
+    import sys
+
+    sys.modules[__name__] = _LazyModule(__name__, globals()["__file__"], _import_structure, module_spec=__spec__)

src/transformers/models/mobilenet_v1/configuration_mobilenet_v1.py

+# coding=utf-8
+# Copyright 2022 The HuggingFace Inc. team. All rights reserved.
+#
+# 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.
+""" MobileNetV1 model configuration"""
+
+from collections import OrderedDict
+from typing import Mapping
+
+from packaging import version
+
+from ...configuration_utils import PretrainedConfig
+from ...onnx import OnnxConfig
+from ...utils import logging
+
+
+logger = logging.get_logger(__name__)
+
+MOBILENET_V1_PRETRAINED_CONFIG_ARCHIVE_MAP = {
+    "google/mobilenet_v1_1.0_224": "https://huggingface.co/google/mobilenet_v1_1.0_224/resolve/main/config.json",
+    "google/mobilenet_v1_0.75_192": "https://huggingface.co/google/mobilenet_v1_0.75_192/resolve/main/config.json",
+    # See all MobileNetV1 models at https://huggingface.co/models?filter=mobilenet_v1
+}
+
+
+class MobileNetV1Config(PretrainedConfig):
+    r"""
+    This is the configuration class to store the configuration of a [`MobileNetV1Model`]. It is used to instantiate a
+    MobileNetV1 model according to the specified arguments, defining the model architecture. Instantiating a
+    configuration with the defaults will yield a similar configuration to that of the MobileNetV1
+    [google/mobilenet_v1_1.0_224](https://huggingface.co/google/mobilenet_v1_1.0_224) architecture.
+
+    Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
+    documentation from [`PretrainedConfig`] for more information.
+
+    Args:
+        num_channels (`int`, *optional*, defaults to 3):
+            The number of input channels.
+        image_size (`int`, *optional*, defaults to 224):
+            The size (resolution) of each image.
+        depth_multiplier (`float`, *optional*, defaults to 1.0):
+            Shrinks or expands the number of channels in each layer. Default is 1.0, which starts the network with 32
+            channels. This is sometimes also called "alpha" or "width multiplier".
+        min_depth (`int`, *optional*, defaults to 8):
+            All layers will have at least this many channels.
+        hidden_act (`str` or `function`, *optional*, defaults to `"relu6"`):
+            The non-linear activation function (function or string) in the Transformer encoder and convolution layers.
+        tf_padding (`bool`, `optional`, defaults to `True`):
+            Whether to use TensorFlow padding rules on the convolution layers.
+        classifier_dropout_prob (`float`, *optional*, defaults to 0.999):
+            The dropout ratio for attached classifiers.
+        initializer_range (`float`, *optional*, defaults to 0.02):
+            The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
+        layer_norm_eps (`float`, *optional*, defaults to 0.001):
+            The epsilon used by the layer normalization layers.
+
+    Example:
+
+    ```python
+    >>> from transformers import MobileNetV1Config, MobileNetV1Model
+
+    >>> # Initializing a "mobilenet_v1_1.0_224" style configuration
+    >>> configuration = MobileNetV1Config()
+
+    >>> # Initializing a model from the "mobilenet_v1_1.0_224" style configuration
+    >>> model = MobileNetV1Model(configuration)
+
+    >>> # Accessing the model configuration
+    >>> configuration = model.config
+    ```"""
+    model_type = "mobilenet_v1"
+
+    def __init__(
+        self,
+        num_channels=3,
+        image_size=224,
+        depth_multiplier=1.0,
+        min_depth=8,
+        hidden_act="relu6",
+        tf_padding=True,
+        classifier_dropout_prob=0.999,
+        initializer_range=0.02,
+        layer_norm_eps=0.001,
+        **kwargs
+    ):
+        super().__init__(**kwargs)
+
+        if depth_multiplier <= 0:
+            raise ValueError("depth_multiplier must be greater than zero.")
+
+        self.num_channels = num_channels
+        self.image_size = image_size
+        self.depth_multiplier = depth_multiplier
+        self.min_depth = min_depth
+        self.hidden_act = hidden_act
+        self.tf_padding = tf_padding
+        self.classifier_dropout_prob = classifier_dropout_prob
+        self.initializer_range = initializer_range
+        self.layer_norm_eps = layer_norm_eps
+
+
+class MobileNetV1OnnxConfig(OnnxConfig):
+
+    torch_onnx_minimum_version = version.parse("1.11")
+
+    @property
+    def inputs(self) -> Mapping[str, Mapping[int, str]]:
+        return OrderedDict([("pixel_values", {0: "batch"})])
+
+    @property
+    def outputs(self) -> Mapping[str, Mapping[int, str]]:
+        if self.task == "image-classification":
+            return OrderedDict([("logits", {0: "batch"})])
+        else:
+            return OrderedDict([("last_hidden_state", {0: "batch"}), ("pooler_output", {0: "batch"})])
+
+    @property
+    def atol_for_validation(self) -> float:
+        return 1e-4

src/transformers/models/mobilenet_v1/convert_original_tf_checkpoint_to_pytorch.py

+# coding=utf-8
+# Copyright 2022 The HuggingFace Inc. team.
+#
+# 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.
+"""Convert MobileNetV1 checkpoints from the tensorflow/models library."""
+
+
+import argparse
+import json
+import re
+from pathlib import Path
+
+import torch
+from PIL import Image
+
+import requests
+from huggingface_hub import hf_hub_download
+from transformers import (
+    MobileNetV1Config,
+    MobileNetV1FeatureExtractor,
+    MobileNetV1ForImageClassification,
+    load_tf_weights_in_mobilenet_v1,
+)
+from transformers.utils import logging
+
+
+logging.set_verbosity_info()
+logger = logging.get_logger(__name__)
+
+
+def get_mobilenet_v1_config(model_name):
+    config = MobileNetV1Config(layer_norm_eps=0.001)
+
+    if "_quant" in model_name:
+        raise ValueError("Quantized models are not supported.")
+
+    matches = re.match(r"^mobilenet_v1_([^_]*)_([^_]*)$", model_name)
+    if matches:
+        config.depth_multiplier = float(matches[1])
+        config.image_size = int(matches[2])
+
+    # The TensorFlow version of MobileNetV1 predicts 1001 classes instead of
+    # the usual 1000. The first class (index 0) is "background".
+    config.num_labels = 1001
+    filename = "imagenet-1k-id2label.json"
+    repo_id = "huggingface/label-files"
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
+    id2label = {int(k) + 1: v for k, v in id2label.items()}
+    id2label[0] = "background"
+    config.id2label = id2label
+    config.label2id = {v: k for k, v in id2label.items()}
+
+    return config
+
+
+# We will verify our results on an image of cute cats
+def prepare_img():
+    url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+    im = Image.open(requests.get(url, stream=True).raw)
+    return im
+
+
+@torch.no_grad()
+def convert_movilevit_checkpoint(model_name, checkpoint_path, pytorch_dump_folder_path, push_to_hub=False):
+    """
+    Copy/paste/tweak model's weights to our MobileNetV1 structure.
+    """
+    config = get_mobilenet_v1_config(model_name)
+
+    # Load 🤗 model
+    model = MobileNetV1ForImageClassification(config).eval()
+
+    # Load weights from TensorFlow checkpoint
+    load_tf_weights_in_mobilenet_v1(model, config, checkpoint_path)
+
+    # Check outputs on an image, prepared by MobileNetV1FeatureExtractor
+    feature_extractor = MobileNetV1FeatureExtractor(
+        crop_size={"width": config.image_size, "height": config.image_size},
+        size={"shortest_edge": config.image_size + 32},
+    )
+    encoding = feature_extractor(images=prepare_img(), return_tensors="pt")
+    outputs = model(**encoding)
+    logits = outputs.logits
+
+    assert logits.shape == (1, 1001)
+
+    if model_name == "mobilenet_v1_1.0_224":
+        expected_logits = torch.tensor([-4.1739, -1.1233, 3.1205])
+    elif model_name == "mobilenet_v1_0.75_192":
+        expected_logits = torch.tensor([-3.9440, -2.3141, -0.3333])
+    else:
+        expected_logits = None
+
+    if expected_logits is not None:
+        assert torch.allclose(logits[0, :3], expected_logits, atol=1e-4)
+
+    Path(pytorch_dump_folder_path).mkdir(exist_ok=True)
+    print(f"Saving model {model_name} to {pytorch_dump_folder_path}")
+    model.save_pretrained(pytorch_dump_folder_path)
+    print(f"Saving feature extractor to {pytorch_dump_folder_path}")
+    feature_extractor.save_pretrained(pytorch_dump_folder_path)
+
+    if push_to_hub:
+        print("Pushing to the hub...")
+        repo_id = "google/" + model_name
+        feature_extractor.push_to_hub(repo_id)
+        model.push_to_hub(repo_id)
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    # Required parameters
+    parser.add_argument(
+        "--model_name",
+        default="mobilenet_v1_1.0_224",
+        type=str,
+        help="Name of the MobileNetV1 model you'd like to convert. Should in the form 'mobilenet_v1_<depth>_<size>'.",
+    )
+    parser.add_argument(
+        "--checkpoint_path", required=True, type=str, help="Path to the original TensorFlow checkpoint (.ckpt file)."
+    )
+    parser.add_argument(
+        "--pytorch_dump_folder_path", required=True, type=str, help="Path to the output PyTorch model directory."
+    )
+    parser.add_argument(
+        "--push_to_hub", action="store_true", help="Whether or not to push the converted model to the 🤗 hub."
+    )
+
+    args = parser.parse_args()
+    convert_movilevit_checkpoint(
+        args.model_name, args.checkpoint_path, args.pytorch_dump_folder_path, args.push_to_hub
+    )

src/transformers/models/mobilenet_v1/feature_extraction_mobilenet_v1.py

+# coding=utf-8
+# Copyright 2022 The HuggingFace Inc. team. All rights reserved.
+#
+# 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.
+"""Feature extractor class for MobileNetV1."""
+
+from ...utils import logging
+from .image_processing_mobilenet_v1 import MobileNetV1ImageProcessor
+
+
+logger = logging.get_logger(__name__)
+
+
+MobileNetV1FeatureExtractor = MobileNetV1ImageProcessor