TAO Pretrained Classification

Description:

TAO Pretrained Classification classifies an image into one of the designated thousands of classes. This model is ready for commercial use.

References:

Other TAO Pre-trained Models

\n
Get TAO Object Detection pre-trained models for YOLOV4, YOLOV3, FasterRCNN, SSD, DSSD, and RetinaNet architectures from the NGC model registry
\n
\n
Get TAO DetectNet_v2 Object Detection pre-trained models for DetectNet_v2 architecture from NGC model registry
\n
\n
Get TAO EfficientDet Object Detection pre-trained models for DetectNet_v2 architecture from NGC model registry
\n
\n
Get TAO Instance segmentation pre-trained models for MaskRCNN architecture from NGC
\n
\n
Get TAO Semantic segmentation pre-trained models for UNet architecture from NGC
\n
\n
Get Purpose-built models from NGC model registry:
\n
- PeopleNet
- TrafficCamNet
- DashCamNet
- FaceDetectIR
- VehicleMakeNet
- VehicleTypeNet
- PeopleSegNet
- PeopleSemSegNet
- License Plate Detection
- License Plate Recognition
- Facial Landmark
- FaceDetect
- 2D Body Pose Net
- ActionRecognitionNet
\n

Model Architecture:

Architecture Type: Convolution Neural Network (CNN)
\nNetwork Architecture(s):

ResNet
VGG
googlenet
mobilenet
squeezenet
darknet
efficientnet
cspdarknet

Input:

Input Type(s): Image
\nInput Format(s): Red, Green, Blue (RGB)
\nInput Parameters: 4D
\nOther Properties Related to Input: 4D includes batch size as fourth dimension. Minimum Resolution: B X 3 X 224 X 224; Maximum Resolution: B X 3 X 518 X 518; No minimum bit depth, alpha, or gamma

Output:

Output Type(s): Label(s)
\nOutput Format: Label: Text String
\nOther Properties Related to Output: Category Label(s): Designated Classes

Software Integration:

Runtime Engine(s):

TAO - 5.2
DeepStream 6.1 or later

Supported Hardware Architecture(s):

Ampere
Jetson
Hopper
Lovelace
Pascal
Turing
Volta

Supported Operating System(s):

Linux
Linux 4 Tegra

Model Version(s):

resnet10/resnet18/resnet34/resnet50/resnet101
vgg16/vgg19
googlenet
mobilenet_v1/mobilenet_v2
squeezenet
darknet19/darknet53
efficientnet_b0
cspdarknet19/cspdarknet53

Training & Evaluation:

Training Dataset:

Link: https://github.com/openimages/dataset/blob/main/READMEV3.md

Data Collection Method by dataset:

Unknown

Labeling Method by dataset:

Unknown

Properties:
\nRoughly 400,000 images and 7,000 validation images across thousands of classes as defined by Google OpenImages Version Three (3) dataset. Most of the human verifications have been done with in-house annotators at Google. A smaller part has been done with crowd-sourced verification from Image Labeler: Crowdsource app, g.co/imagelabeler.

Evaluation Dataset:

Link: https://github.com/openimages/dataset/blob/main/READMEV3.md

Data Collection Method by dataset:

Unknown

Labeling Method by dataset:

Unknown

Properties:
\n15,000 test images from Google OpenImages Version Three (3) dataset.

Running Image classification Models Using TAO

The image classification apps in TAO expect data in KITTI file format. TAO provides a simple command line interface to train a deep learning model for image classification.

The models in this model area are only compatible with TAO Toolkit. For more information about TAO please refer to the Getting Started Guide.

\n
Install the NGC CLI from ngc.nvidia.com
\n
\n
Configure the NGC CLI using the following command
\n

ngc config set\n

To view all the backbones for classification that are supported by TAO:

ngc registry model list nvidia/tao_pretrained_classification:*\n

Download the model:

--dest \n\">

ngc registry model download-version nvidia/tao_pretrained_classification:<template> --dest <path>\n

Instructions to run the sample notebook

\n
Get the NGC API key from the SETUP tab on the left. Please store this key for future use. Detailed instructions can be found here
\n
\n
Configure the NGC command line interface using the command mentioned below and follow the prompts.
\n

ngc config set\n

Download the sample notebooks from NGC using the command below

ngc registry resource download-version \"nvidia/tao_cv_samples:v1.0.2\"\n

Invoke the jupyter notebook using the following command

jupyter notebook --ip 0.0.0.0 --port 8888 --allow-root\n

Open an internet browser and type in the following URL to start running the notebooks when running on a local machine.

http://0.0.0.0:8888\n

If you wish to run view the notebook from a remote client, please modify the URL as follows:

http://a.b.c.d:8888\n

Where, the a.b.c.d is the IP address of the machine running the container.

Technical blogs

Access the latest in Vision AI development workflows with NVIDIA TAO Toolkit 5.0
Improve accuracy and robustness of vision ai models with vision transformers and NVIDIA TAO
Train like a ‘pro’ without being an AI expert using TAO AutoML
Create Custom AI models using NVIDIA TAO Toolkit with Azure Machine Learning
Developing and Deploying AI-powered Robots with NVIDIA Isaac Sim and NVIDIA TAO
Learn endless ways to adapt and supercharge your AI workflows with TAO - Whitepaper
Customize Action Recognition with TAO and deploy with DeepStream
Read the 2 part blog on training and optimizing 2D body pose estimation model with TAO - Part 1 | Part 2
Learn how to train real-time License plate detection and recognition app with TAO and DeepStream
Model accuracy is extremely important, learn how you can achieve state of the art accuracy for classification and object detection models using TAO

Ethical Considerations:

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse. For more detailed information on ethical considerations for this model, please see the Model Card++ Promise and the Explainability, Bias, Safety & Security, and Privacy Subcards.

","resourceVersions",[190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208],"resourceVersionsTotal",18,"hasMoreVersions","versionMetadata",{"_164":21,"_165":166,"_48":167,"_168":-5,"_169":170,"_171":172,"_173":79,"_174":16,"_175":30,"_176":-5,"_177":30,"_178":-5,"_179":30,"_180":-5,"_181":30,"_182":-5,"_183":30,"_184":-5,"_185":30,"_186":-5,"_187":30,"_188":189},"versionNotFound","isLoggedIn","latestVersionPath","/orgs/nvidia/tao/models/pretrained_classification/-","version","imageReference","cliDownloadCommand","ngc registry model download-version nvidia/tao/pretrained_classification:mobilenet_v2","helmFetchCommand","viewAllVersionsPath","/orgs/nvidia/tao/models/pretrained_classification/mobilenet_v2/version-history","accessCheckResult",{"_163":85},"hasRelatedCollections","context",{"_159":160,"_161":162},[],[],[153,154,155,156,157,158],{"_44":58,"_46":69,"_48":60,"_50":61,"_52":70,"_54":55},{"_44":58,"_46":67,"_48":60,"_50":61,"_52":68,"_54":55},{"_44":58,"_46":65,"_48":60,"_50":61,"_52":66,"_54":55},{"_44":58,"_46":59,"_48":60,"_50":61,"_52":62,"_54":55},{"_44":56,"_46":57,"_48":101,"_50":-5,"_52":-5,"_54":55},{"_44":45,"_46":47,"_48":49,"_50":51,"_52":53,"_54":55},"orgName","nvidia","teamName","tao","allowed","versionId","description","","UPLOAD_COMPLETE","releaseType","createdDate",["D",1629317867567],"updatedDate",["D",1740840628891],"totalSizeInBytes","totalFileCount","isSigned","malwareScanStatus","isScanned","releaseNotes","hasReleaseNotes","setup","hasSetup","quickStartGuide","hasQuickStartGuide","advanced","hasAdvanced","performance","hasPerformance","policies",[],{"_75":244,"_77":245,"_168":-5,"_175":30,"_177":30,"_188":243,"_179":30},{"_75":242,"_77":242,"_168":-5,"_175":30,"_177":30,"_188":243,"_179":30},{"_75":240,"_77":240,"_168":-5,"_175":30,"_177":30,"_188":241,"_179":30},{"_75":238,"_77":238,"_168":-5,"_175":30,"_177":30,"_188":239,"_179":30},{"_75":236,"_77":236,"_168":-5,"_175":30,"_177":30,"_188":237,"_179":30},{"_75":234,"_77":234,"_168":-5,"_175":30,"_177":30,"_188":235,"_179":30},{"_75":21,"_77":21,"_168":-5,"_175":30,"_177":30,"_188":233,"_179":30},{"_75":231,"_77":231,"_168":-5,"_175":30,"_177":30,"_188":232,"_179":30},{"_75":229,"_77":229,"_168":-5,"_175":30,"_177":30,"_188":230,"_179":30},{"_75":227,"_77":227,"_168":-5,"_175":30,"_177":30,"_188":228,"_179":30},{"_75":225,"_77":225,"_168":-5,"_175":30,"_177":30,"_188":226,"_179":30},{"_75":223,"_77":223,"_168":-5,"_175":30,"_177":30,"_188":224,"_179":30},{"_75":221,"_77":221,"_168":-5,"_175":30,"_177":30,"_188":222,"_179":30},{"_75":219,"_77":219,"_168":-5,"_175":30,"_177":30,"_188":220,"_179":30},{"_75":217,"_77":217,"_168":-5,"_175":30,"_177":30,"_188":218,"_179":30},{"_75":215,"_77":215,"_168":-5,"_175":30,"_177":30,"_188":216,"_179":30},{"_75":213,"_77":213,"_168":-5,"_175":30,"_177":30,"_188":214,"_179":30},{"_75":211,"_77":211,"_168":-5,"_175":30,"_177":30,"_188":212,"_179":30},{"_75":209,"_77":209,"_168":-5,"_175":30,"_177":30,"_188":210,"_179":30},"resnet10",[],"efficientnet_b1_relu",[],"resnet18",[],"efficientnet_b1_swish",[],"resnet34",[],"resnet50",[],"resnet101",[],"vgg16",[],"vgg19",[],"darknet53",[],"darknet19",[],"mobilenet_v1",[],[],"squeezenet",[],"googlenet",[],"cspdarknet19",[],"cspdarknet53",[],"cspdarknet_tiny",[],"-","Latest (cspdarknet_tiny)","nvidia/tao/pretrained_classification","key","nvidia/tao/models/pretrained_classification","resourceType","models","resourceId","pretrained_classification","shortDescription","Pretrained weights to facilitate transfer learning using TAO Toolkit.","# TAO Pretrained Classification\n\n## Description:\nTAO Pretrained Classification classifies an image into one of the designated thousands of classes. This model is ready for commercial use.\n\n## References:\n### Other TAO Pre-trained Models\n\n- Get [TAO Object Detection](https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_object_detection) pre-trained models for **YOLOV4, YOLOV3, FasterRCNN, SSD, DSSD, and RetinaNet** architectures from the NGC model registry\n- Get [TAO DetectNet_v2 Object Detection](https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_detectnet_v2) pre-trained models for **DetectNet_v2** architecture from NGC model registry\n- Get [TAO EfficientDet Object Detection](https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_efficientdet) pre-trained models for **DetectNet_v2** architecture from NGC model registry\n- Get [TAO Instance segmentation](https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_instance_segmentation) pre-trained models for **MaskRCNN** architecture from NGC\n- Get [TAO Semantic segmentation](https://ngc.nvidia.com/catalog/models/nvidia:tao:pretrained_semantic_segmentation) pre-trained models for **UNet** architecture from NGC\n\n- Get Purpose-built models from NGC model registry:\n - [PeopleNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:peoplenet)\n - [TrafficCamNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:trafficcamnet)\n - [DashCamNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:dashcamnet)\n - [FaceDetectIR](https://ngc.nvidia.com/catalog/models/nvidia:tao:facedetectir)\n - [VehicleMakeNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:vehiclemakenet)\n - [VehicleTypeNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:vehicletypenet)\n - [PeopleSegNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:peoplesegnet)\n - [PeopleSemSegNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:peoplesemsegnet)\n - [License Plate Detection](https://ngc.nvidia.com/catalog/models/nvidia:tao:lpdnet)\n - [License Plate Recognition](https://ngc.nvidia.com/catalog/models/nvidia:tao:lprnet)\n - [Facial Landmark](https://ngc.nvidia.com/catalog/models/nvidia:tao:fpenet)\n - [FaceDetect](https://ngc.nvidia.com/catalog/models/nvidia:facenet)\n - [2D Body Pose Net](https://ngc.nvidia.com/catalog/models/nvidia:tao:bodyposenet)\n - [ActionRecognitionNet](https://ngc.nvidia.com/catalog/models/nvidia:tao:actionrecognitionnet)\n\n## Model Architecture: \n**Architecture Type:** Convolution Neural Network (CNN)
\n**Network Architecture(s):**
\n- ResNet\n- VGG\n- googlenet\n- mobilenet\n- squeezenet\n- darknet\n- efficientnet\n- cspdarknet\n\n## Input:\n**Input Type(s):** Image
\n**Input Format(s):** Red, Green, Blue (RGB)
\n**Input Parameters:** 4D
\n**Other Properties Related to Input:** 4D includes batch size as fourth dimension. Minimum Resolution: B X 3 X 224 X 224; Maximum Resolution: B X 3 X 518 X 518; No minimum bit depth, alpha, or gamma
\n\n## Output:\n**Output Type(s):** Label(s)
\n**Output Format:** Label: Text String
\n**Other Properties Related to Output:** Category Label(s): Designated Classes
\n\n## Software Integration:\n**Runtime Engine(s):** \n* TAO - 5.2
\n* DeepStream 6.1 or later
\n\n**Supported Hardware Architecture(s):**
\n* Ampere
\n* Jetson
\n* Hopper
\n* Lovelace
\n* Pascal
\n* Turing
\n* Volta
\n\n**Supported Operating System(s):**
\n* Linux
\n* Linux 4 Tegra
\n\n## Model Version(s): \n- resnet10/resnet18/resnet34/resnet50/resnet101\n- vgg16/vgg19\n- googlenet\n- mobilenet_v1/mobilenet_v2\n- squeezenet\n- darknet19/darknet53\n- efficientnet_b0\n- cspdarknet19/cspdarknet53\n\n## Training & Evaluation: \n\n## Training Dataset:\n**Link:** https://github.com/openimages/dataset/blob/main/READMEV3.md
\n\n**Data Collection Method by dataset:**
\n* Unknown
\n\n**Labeling Method by dataset:**
\n* Unknown
\n\n**Properties:**
\nRoughly 400,000 images and 7,000 validation images across thousands of classes as defined by [Google OpenImages Version Three (3) dataset](https://github.com/openimages/dataset/blob/main/READMEV3.md). Most of the human verifications have been done with in-house annotators at Google. A smaller part has been done with crowd-sourced verification from Image Labeler: [Crowdsource app](http://play.google.com/store/apps/details?id=com.google.android.apps.village.boond), [g.co/imagelabeler](http://g.co/imagelabeler).\n\n\n## Evaluation Dataset:\n**Link:** https://github.com/openimages/dataset/blob/main/READMEV3.md
\n\n**Data Collection Method by dataset:**
\n* Unknown
\n\n**Labeling Method by dataset:**
\n* Unknown
\n\n**Properties:**
\n15,000 test images from [Google OpenImages Version Three (3) dataset](https://github.com/openimages/dataset/blob/main/READMEV3.md).\n\n## Running Image classification Models Using TAO \n\nThe image classification apps in TAO expect data in KITTI file format. TAO provides a simple command line interface to train a deep learning model for image classification.\n\nThe models in this model area are only compatible with TAO Toolkit. For more information about TAO please refer to the [Getting Started Guide](https://docs.nvidia.com/tao/tao-toolkit/text/overview.html).\n\n1. Install the NGC CLI from `ngc.nvidia.com`\n\n2. Configure the NGC CLI using the following command\n\n ```sh\n ngc config set\n ```\n\n3. To view all the backbones for classification that are supported by TAO:\n\n ```sh\n ngc registry model list nvidia/tao_pretrained_classification:*\n ```\n\n4. Download the model:\n\n ```sh\n ngc registry model download-version nvidia/tao_pretrained_classification: