Model Zoo

November 6, 2023 · View on GitHub

Computer Vision Models - Pretrained Checkpoints

You can load any of our pretrained model in 2 lines of code:

from super_gradients.training import models
from super_gradients.common.object_names import Models

model = models.get(Models.YOLOX_S, pretrained_weights="coco")

All the available models are listed in the column Model name.

Pretrained Classification PyTorch Checkpoints

Model	Model name	Dataset	Resolution	Top-1	Top-5	Latency (HW)*_T4	Latency (Production)**_T4	Latency (HW)*_{Jetson Xavier NX}	Latency (Production)**_{Jetson Xavier NX}	Latency _{Cascade Lake}	Torch Compile Support
ViT base	vit_base	ImageNet21K	224x224	84.15	-	4.46ms	4.60ms	- *	-	57.22ms	Not Supported
ViT large	vit_large	ImageNet21K	224x224	85.64	-	12.81ms	13.19ms	- *	-	187.22ms	Not Supported
BEiT	beit_base_patch16_224	ImageNet21K	224x224	-	-	-ms	-ms	- *	-	-ms	Supported
EfficientNet B0	efficientnet_b0	ImageNet	224x224	77.62	93.49	0.93ms	1.38ms	- *	-	3.44ms	Supported
RegNet Y200	regnetY200	ImageNet	224x224	70.88	89.35	0.63ms	1.08ms	2.16ms	2.47ms	2.06ms	Supported
RegNet Y400	regnetY400	ImageNet	224x224	74.74	91.46	0.80ms	1.25ms	2.62ms	2.91ms	2.87ms	Supported
RegNet Y600	regnetY600	ImageNet	224x224	76.18	92.34	0.77ms	1.22ms	2.64ms	2.93ms	2.39ms	Supported
RegNet Y800	regnetY800	ImageNet	224x224	77.07	93.26	0.74ms	1.19ms	2.77ms	3.04ms	2.81ms	Supported
ResNet 18	resnet18	ImageNet	224x224	70.6	89.64	0.52ms	0.95ms	2.01ms	2.30ms	4.56ms	Supported
ResNet 34	resnet34	ImageNet	224x224	74.13	91.7	0.92ms	1.34ms	3.57ms	3.87ms	7.64ms	Supported
ResNet 50	resnet50	ImageNet	224x224	81.91	93.0	1.03ms	1.44ms	4.78ms	5.10ms	9.25ms	Supported
MobileNet V3_large-300 epochs	mobilenet_v3_large	ImageNet	224x224	74.52	91.92	0.67ms	1.11ms	2.42ms	2.71ms	1.76ms	Supported
MobileNet V3_small	mobilenet_v3_small	ImageNet	224x224	67.45	87.47	0.55ms	0.96ms	2.01ms *	2.35ms	1.06ms	Supported
MobileNet V2_w1	mobilenet_v2	ImageNet	224x224	73.08	91.1	0.46 ms	0.89ms	1.65ms *	1.90ms	1.56ms	Supported

NOTE:

Latency (HW)* - Hardware performance (not including IO)

Latency (Production)** - Production Performance (including IO)

Performance measured for T4 and Jetson Xavier NX with TensorRT, using FP16 precision and batch size 1

Performance measured for Cascade Lake CPU with OpenVINO, using FP16 precision and batch size 1

Pretrained Object Detection PyTorch Checkpoints

Model	Model Name	Dataset	Resolution	mAP^val 0.5:0.95	Latency (HW)*_T4	Latency (Production)**_T4	Latency (HW)*_{Jetson Xavier NX}	Latency (Production)**_{Jetson Xavier NX}	Latency _{Cascade Lake}	Torch Compile Support
YOLO-NAS S	yolo_nas_s	COCO	640x640	47.5(FP16) 47.03(INT8)	3.21(FP16) 2.36(INT8)					Supported
YOLO-NAS M	yolo_nas_m	COCO	640x640	51.55(FP16) 51.0(INT8)	5.85(FP16) 3.78(INT8)					Supported
YOLO-NAS L	yolo_nas_l	COCO	640x640	52.22(FP16) 52.1(INT8)	7.87(FP16) 4.78(INT8)					Supported
PP-YOLOE small	ppyoloe_s	COCO	640x640	42.52	2.39ms	4.3ms	14.28ms	14.99ms	-	Not Supported
PP-YOLOE medium	ppyoloe_m	COCO	640x640	47.11	5.16ms	7.05ms	32.71ms	33.46ms	-	Not Supported
PP-YOLOE large	ppyoloe_l	COCO	640x640	49.48	7.65ms	9.59ms	51.13ms	50.39ms	-	Not Supported
PP-YOLOE x-large	ppyoloe_x	COCO	640x640	51.15	14.04ms	15.96ms	94.92ms	94.22ms	-	Not Supported
YOLOX nano	yolox_n	COCO	640x640	26.77	2.47ms	4.09ms	11.49ms	12.97ms	-	Not Supported
YOLOX tiny	yolox_t	COCO	640x640	37.18	3.16ms	4.61ms	15.23ms	19.24ms	-	Not Supported
YOLOX small	yolox_s	COCO	640x640	40.47	3.58ms	4.94ms	18.88ms	22.48ms	-	Not Supported
YOLOX medium	yolox_m	COCO	640x640	46.4	6.40ms	7.65ms	39.22ms	44.5ms	-	Not Supported
YOLOX large	yolox_l	COCO	640x640	49.25	10.07ms	11.12ms	68.73ms	77.01ms	-	Not Supported
SSD lite MobileNet v2	ssd_lite_mobilenet_v2	COCO	320x320	21.5	0.77ms	1.40ms	5.28ms	6.44ms	4.13ms	Not Supported
SSD lite MobileNet v1	ssd_mobilenet_v1	COCO	320x320	24.3	1.55ms	2.84ms	8.07ms	9.14ms	22.76ms	Not Supported

NOTE:

Latency (HW)* - Hardware performance (not including IO)

Latency (Production)** - Production Performance (including IO)

Latency performance measured for T4 and Jetson Xavier NX with TensorRT, using FP16 precision and batch size 1

Latency performance measured for Cascade Lake CPU with OpenVINO, using FP16 precision and batch size 1

Pretrained Semantic Segmentation PyTorch Checkpoints

Model	Model Name	Dataset	Resolution	mIoU	Latency b1_T4	Latency b1_T4 including IO	Latency (Production)**_{Jetson Xavier NX}	Torch Compile Support
PP-LiteSeg B50	pp_lite_b_seg50	Cityscapes	512x1024	76.48	4.18ms	31.22ms	31.69ms	Supported
PP-LiteSeg B75	pp_lite_b_seg75	Cityscapes	768x1536	78.52	6.84ms	33.69ms	49.89ms	Supported
PP-LiteSeg T50	pp_lite_t_seg50	Cityscapes	512x1024	74.92	3.26ms	30.33ms	26.20ms	Supported
PP-LiteSeg T75	pp_lite_t_seg75	Cityscapes	768x1536	77.56	5.20ms	32.28ms	38.03ms	Supported
DDRNet 23 slim	ddrnet_23_slim	Cityscapes	1024x2048	79.41	5.74ms	32.01ms	45.18ms	Supported
DDRNet 23	ddrnet_23	Cityscapes	1024x2048	81.48	12.74ms	39.01ms	106.26ms	Supported
DDRNet 39	ddrnet_39	Cityscapes	1024x2048	81.32	23.57ms	52.41ms	145.79ms	Supported
STDC 1-Seg50	stdc1_seg50	Cityscapes	512x1024	75.11	3.34ms	30.12ms	27.54ms	Supported
STDC 1-Seg75	stdc1_seg75	Cityscapes	768x1536	77.8	5.53ms	32.490ms	43.88	Supported
STDC 2-Seg50	stdc2_seg50	Cityscapes	512x1024	76.44	4.12ms	30.94ms	32.03ms	Supported
STDC 2-Seg75	stdc2_seg75	Cityscapes	768x1536	78.93	6.95ms	33.89ms	54.48ms	Supported
RegSeg (exp48)	regseg48	Cityscapes	1024x2048	78.15	12.03ms	38.91ms	78.20ms	Supported

NOTE:

Performance measured on T4 GPU with TensorRT, using FP16 precision and batch size 1 (latency), and not including IO

For resolutions below 1024x2048 we first resize the input to the inference resolution and then resize the predictions to 1024x2048. The time of resizing is included in the measurements so that the practical input-size is 1024x2048.

DDRNet23 and DDRNet23_Slim results were achieved with channel wise knowledge distillation training recipe.

Pretrained Pose Estimation PyTorch Checkpoints

Model	Model Name	Dataset	Resolution	AP (No TTA / H-Flip TTA / H-Flip TTA+Rescoring)	Latency b1_T4	Latency b1_T4 including IO	Latency (Production)**_{Jetson Xavier NX}
DEKR_W32_NO_DC	dekr_w32_no_dc	COCO2017 PE	640x640	63.08 / 64.96 / 67.32	13.29 ms	15.31 ms	75.99 ms
YoloNAS POSE N	yolo_nas_pose_n	COCO2017 PE	640x640	59.68 / N/A / N/A	N/A	2.35 ms	15.99 ms
YoloNAS POSE S	yolo_nas_pose_s	COCO2017 PE	640x640	64.15 / N/A / N/A	N/A	3.29 ms	21.01 ms
YoloNAS POSE M	yolo_nas_pose_m	COCO2017 PE	640x640	67.87 / N/A / N/A	N/A	6.87 ms	38.40 ms
YoloNAS POSE L	yolo_nas_pose_l	COCO2017 PE	640x640	68.24 / N/A / N/A	N/A	8.86 ms	49.34 ms

Implemented Model Architectures

Image Classification

DensNet (Densely Connected Convolutional Networks) - Densely Connected Convolutional Networks https://arxiv.org/pdf/1608.06993.pdf
DPN - Dual Path Networks https://arxiv.org/pdf/1707.01629
EfficientNet - https://arxiv.org/abs/1905.11946
GoogleNet - https://arxiv.org/pdf/1409.4842
LeNet - https://yann.lecun.com/exdb/lenet/
MobileNet - Efficient Convolutional Neural Networks for Mobile Vision Applications https://arxiv.org/pdf/1704.04861
MobileNet v2 - https://arxiv.org/pdf/1801.04381
MobileNet v3 - https://arxiv.org/pdf/1905.02244
PNASNet - Progressive Neural Architecture Search Networks https://arxiv.org/pdf/1712.00559
Pre-activation ResNet - https://arxiv.org/pdf/1603.05027
RegNet - https://arxiv.org/pdf/2003.13678.pdf
RepVGG - Making VGG-style ConvNets Great Again https://arxiv.org/pdf/2101.03697.pdf
ResNet - Deep Residual Learning for Image Recognition https://arxiv.org/pdf/1512.03385
ResNeXt - Aggregated Residual Transformations for Deep Neural Networks https://arxiv.org/pdf/1611.05431
SENet - Squeeze-and-Excitation Networkshttps://arxiv.org/pdf/1709.01507
ShuffleNet - https://arxiv.org/pdf/1707.01083
ShuffleNet v2 - Efficient Convolutional Neural Network for Mobile Deviceshttps://arxiv.org/pdf/1807.11164
VGG - Very Deep Convolutional Networks for Large-scale Image Recognition https://arxiv.org/pdf/1409.1556

Object Detection

Semantic Segmentation

PP-LiteSeg - https://arxiv.org/pdf/2204.02681v1.pdf
DDRNet (Deep Dual-resolution Networks) - https://arxiv.org/pdf/2101.06085.pdf
LadderNet - Multi-path networks based on U-Net for medical image segmentation https://arxiv.org/pdf/1810.07810
RegSeg - Rethink Dilated Convolution for Real-time Semantic Segmentation https://arxiv.org/pdf/2111.09957
ShelfNet - https://arxiv.org/pdf/1811.11254
STDC - Rethinking BiSeNet For Real-time Semantic Segmentation https://arxiv.org/pdf/2104.13188

Pose Estimation

HRNet DEKR - Bottom-Up Human Pose Estimation Via Disentangled Keypoint Regression https://arxiv.org/pdf/2104.02300.pdf
YoloNAS Pose