diff --git a/README.md b/README.md index b9436cb2082b660280069df8e96274dca82799a8..d15f9dc95f08f0155e40d6a2be19bacea0fc3738 100644 --- a/README.md +++ b/README.md @@ -14,1425 +14,229 @@ DeepSparkInference将按季度进行版本更新，后续会逐步丰富模型 ### Classification - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
AlexNetFP16SupportedSupported
INT8SupportedSupported
CLIPFP16SupportedIxFormer
INT8--
Conformer-BFP16Supported-
INT8--
ConvNeXt-BaseFP16SupportedSupported
INT8--
ConvNext-S (OpenMMLab)FP16Supported-
INT8--
ConvNeXt-SmallFP16SupportedSupported
INT8--
CSPDarkNet53FP16SupportedSupported
INT8-Supported
CSPResNet50FP16SupportedSupported
INT8-Supported
DeiT-tinyFP16SupportedSupported
INT8--
DenseNet121FP16SupportedSupported
INT8--
DenseNet161FP16SupportedSupported
INT8--
DenseNet169FP16SupportedSupported
INT8--
DenseNet201FP16SupportedSupported
INT8--
EfficientNet-B0FP16SupportedSupported
INT8-Supported
EfficientNet-B1FP16SupportedSupported
INT8-Supported
EfficientNet-B2FP16SupportedSupported
INT8--
EfficientNet-B3FP16SupportedSupported
INT8--
EfficientNet-B4FP16Supported-
INT8--
EfficientNetV2FP16SupportedSupported
INT8-Supported
EfficientNetv2_rw_tFP16SupportedSupported
INT8--
EfficientNetv2_sFP16SupportedSupported
INT8--
GoogLeNetFP16SupportedSupported
INT8SupportedSupported
HRNet-W18FP16SupportedSupported
INT8-Supported
InceptionV3FP16SupportedSupported
INT8SupportedSupported
Inception_ResNet_V2FP16-Supported
INT8-Supported
Mixer_BFP16Supported-
INT8--
MNASNet0_5FP16Supported-
INT8--
MNASNet0_75FP16Supported-
INT8--
MobileNetV2FP16SupportedSupported
INT8SupportedSupported
MobileNetV3_LargeFP16Supported-
INT8--
MobileNetV3_SmallFP16SupportedSupported
INT8--
MViTv2_baseFP16Supported-
INT8--
RegNet_x_16gfFP16Supported-
INT8--
RegNet_x_1_6gfFP16Supported-
INT8--
RegNet_y_1_6gfFP16Supported-
INT8--
RepVGGFP16SupportedSupported
INT8--
Res2Net50FP16SupportedSupported
INT8-Supported
ResNeSt50FP16Supported-
INT8--
ResNet101FP16SupportedSupported
INT8SupportedSupported
ResNet152FP16Supported-
INT8Supported-
ResNet18FP16SupportedSupported
INT8SupportedSupported
ResNet34FP16-Supported
INT8-Supported
ResNet50FP16SupportedSupported
INT8Supported-
ResNetV1D50FP16SupportedSupported
INT8-Supported
ResNeXt50_32x4dFP16SupportedSupported
INT8--
ResNeXt101_64x4dFP16Supported-
INT8--
ResNeXt101_32x8dFP16Supported-
INT8--
SEResNet50FP16Supported-
INT8--
ShuffleNetV1FP16-Supported
INT8--
ShuffleNetV2_x0_5FP16Supported-
INT8--
ShuffleNetV2_x1_0FP16Supported-
INT8--
ShuffleNetV2_x1_5FP16Supported-
INT8--
ShuffleNetV2_x2_0FP16Supported-
INT8--
SqueezeNet 1.0FP16SupportedSupported
INT8-Supported
SqueezeNet 1.1FP16-Supported
INT8-Supported
SVT BaseFP16Supported-
INT8--
Swin TransformerFP16Supported-
INT8--
Swin Transformer LargeFP16-Supported
INT8--
VGG11FP16Supported-
INT8--
VGG16FP16SupportedSupported
INT8Supported-
Wide ResNet50FP16SupportedSupported
INT8SupportedSupported
Wide ResNet101FP16Supported-
INT8--
+| Model | Prec. | IGIE | IxRT | +|------------------------|-------|-----------------------------------------------------------------------|-----------------------------------------------------------------------| +| AlexNet | FP16 | [✅](models/cv/classification/alexnet/igie/README.md#fp16) | [✅](models/cv/classification/alexnet/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/alexnet/igie/README.md#int8) | [✅](models/cv/classification/alexnet/ixrt/README.md#int8) | +| CLIP | FP16 | [✅](models/cv/classification/clip/igie/README.md#fp16) | [✅](models/cv/classification/clip/ixformer/README.md#fp16) | +| Conformer-B | FP16 | [✅](models/cv/classification/conformer_base/igie/README.md#fp16) | | +| ConvNeXt-Base | FP16 | [✅](models/cv/classification/convnext_base/igie/README.md#fp16) | [✅](models/cv/classification/convnext_base/ixrt/README.md#fp16) | +| ConvNext-S | FP16 | [✅](models/cv/classification/convnext_s/igie/README.md#fp16) | | +| ConvNeXt-Small | FP16 | [✅](models/cv/classification/convnext_small/igie/README.md#fp16) | [✅](models/cv/classification/convnext_small/ixrt/README.md#fp16) | +| CSPDarkNet53 | FP16 | [✅](models/cv/classification/cspdarknet53/igie/README.md#fp16) | [✅](models/cv/classification/cspdarknet53/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/cspdarknet53/ixrt/README.md#int8) | +| CSPResNet50 | FP16 | [✅](models/cv/classification/cspresnet50/igie/README.md#fp16) | [✅](models/cv/classification/cspresnet50/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/cspresnet50/ixrt/README.md#int8) | +| DeiT-tiny | FP16 | [✅](models/cv/classification/deit_tiny/igie/README.md#fp16) | [✅](models/cv/classification/deit_tiny/ixrt/README.md#fp16) | +| DenseNet121 | FP16 | [✅](models/cv/classification/densenet121/igie/README.md#fp16) | [✅](models/cv/classification/densenet121/ixrt/README.md#fp16) | +| DenseNet161 | FP16 | [✅](models/cv/classification/densenet161/igie/README.md#fp16) | [✅](models/cv/classification/densenet161/ixrt/README.md#fp16) | +| DenseNet169 | FP16 | [✅](models/cv/classification/densenet169/igie/README.md#fp16) | [✅](models/cv/classification/densenet169/ixrt/README.md#fp16) | +| DenseNet201 | FP16 | [✅](models/cv/classification/densenet201/igie/README.md#fp16) | [✅](models/cv/classification/densenet201/ixrt/README.md#fp16) | +| EfficientNet-B0 | FP16 | [✅](models/cv/classification/efficientnet_b0/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_b0/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/efficientnet_b0/ixrt/README.md#int8) | +| EfficientNet-B1 | FP16 | [✅](models/cv/classification/efficientnet_b1/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_b1/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/efficientnet_b1/ixrt/README.md#int8) | +| EfficientNet-B2 | FP16 | [✅](models/cv/classification/efficientnet_b2/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_b2/ixrt/README.md#fp16) | +| EfficientNet-B3 | FP16 | [✅](models/cv/classification/efficientnet_b3/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_b3/ixrt/README.md#fp16) | +| EfficientNet-B4 | FP16 | [✅](models/cv/classification/efficientnet_b4/igie/README.md#fp16) | | +| EfficientNetV2 | FP16 | [✅](models/cv/classification/efficientnet_v2/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_v2/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/efficientnet_v2/ixrt/README.md#int8) | +| EfficientNetv2_rw_t | FP16 | [✅](models/cv/classification/efficientnetv2_rw_t/igie/README.md#fp16) | [✅](models/cv/classification/efficientnetv2_rw_t/ixrt/README.md#fp16) | +| EfficientNetv2_s | FP16 | [✅](models/cv/classification/efficientnet_v2_s/igie/README.md#fp16) | [✅](models/cv/classification/efficientnet_v2_s/ixrt/README.md#fp16) | +| GoogLeNet | FP16 | [✅](models/cv/classification/googlenet/igie/README.md#fp16) | [✅](models/cv/classification/googlenet/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/googlenet/igie/README.md#int8) | [✅](models/cv/classification/googlenet/ixrt/README.md#int8) | +| HRNet-W18 | FP16 | [✅](models/cv/classification/hrnet_w18/igie/README.md#fp16) | [✅](models/cv/classification/hrnet_w18/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/hrnet_w18/ixrt/README.md#int8) | +| InceptionV3 | FP16 | [✅](models/cv/classification/inception_v3/igie/README.md#fp16) | [✅](models/cv/classification/inception_v3/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/inception_v3/igie/README.md#int8) | [✅](models/cv/classification/inception_v3/ixrt/README.md#int8) | +| Inception_ResNet_V2 | FP16 | | [✅](models/cv/classification/inceptionresnetv2/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/inceptionresnetv2/ixrt/README.md#int8) | +| Mixer_B | FP16 | [✅](models/cv/classification/mlp_mixer_base/igie/README.md#fp16) | | +| MNASNet0_5 | FP16 | [✅](models/cv/classification/mnasnet0_5/igie/README.md#fp16) | | +| MNASNet0_75 | FP16 | [✅](models/cv/classification/mnasnet0_75/igie/README.md#fp16) | | +| MobileNetV2 | FP16 | [✅](models/cv/classification/mobilenet_v2/igie/README.md#fp16) | [✅](models/cv/classification/mobilenet_v2/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/mobilenet_v2/igie/README.md#int8) | [✅](models/cv/classification/mobilenet_v2/ixrt/README.md#int8) | +| MobileNetV3_Large | FP16 | [✅](models/cv/classification/mobilenet_v3_large/igie/README.md#fp16) | | +| MobileNetV3_Small | FP16 | [✅](models/cv/classification/mobilenet_v3/igie/README.md#fp16) | [✅](models/cv/classification/mobilenet_v3/ixrt/README.md#fp16) | +| MViTv2_base | FP16 | [✅](models/cv/classification/mvitv2_base/igie/README.md#fp16) | | +| RegNet_x_16gf | FP16 | [✅](models/cv/classification/regnet_x_16gf/igie/README.md#fp16) | | +| RegNet_x_1_6gf | FP16 | [✅](models/cv/classification/regnet_x_1_6gf/igie/README.md#fp16) | | +| RegNet_y_1_6gf | FP16 | [✅](models/cv/classification/regnet_y_1_6gf/igie/README.md#fp16) | | +| RepVGG | FP16 | [✅](models/cv/classification/repvgg/igie/README.md#fp16) | [✅](models/cv/classification/repvgg/ixrt/README.md#fp16) | +| Res2Net50 | FP16 | [✅](models/cv/classification/res2net50/igie/README.md#fp16) | [✅](models/cv/classification/res2net50/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/res2net50/ixrt/README.md#int8) | +| ResNeSt50 | FP16 | [✅](models/cv/classification/resnest50/igie/README.md#fp16) | | +| ResNet101 | FP16 | [✅](models/cv/classification/resnet101/igie/README.md#fp16) | [✅](models/cv/classification/resnet101/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/resnet101/igie/README.md#int8) | [✅](models/cv/classification/resnet101/ixrt/README.md#int8) | +| ResNet152 | FP16 | [✅](models/cv/classification/resnet152/igie/README.md#fp16) | | +| | INT8 | [✅](models/cv/classification/resnet152/igie/README.md#int8) | | +| ResNet18 | FP16 | [✅](models/cv/classification/resnet18/igie/README.md#fp16) | [✅](models/cv/classification/resnet18/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/resnet18/igie/README.md#int8) | [✅](models/cv/classification/resnet18/ixrt/README.md#int8) | +| ResNet34 | FP16 | | [✅](models/cv/classification/resnet34/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/resnet34/ixrt/README.md#int8) | +| ResNet50 | FP16 | [✅](models/cv/classification/resnet50/igie/README.md#fp16) | [✅](models/cv/classification/resnet50/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/resnet50/igie/README.md#int8) | | +| ResNetV1D50 | FP16 | [✅](models/cv/classification/resnetv1d50/igie/README.md#fp16) | [✅](models/cv/classification/resnetv1d50/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/resnetv1d50/ixrt/README.md#int8) | +| ResNeXt50_32x4d | FP16 | [✅](models/cv/classification/resnext50_32x4d/igie/README.md#fp16) | [✅](models/cv/classification/resnext50_32x4d/ixrt/README.md#fp16) | +| ResNeXt101_64x4d | FP16 | [✅](models/cv/classification/resnext101_64x4d/igie/README.md#fp16) | | +| ResNeXt101_32x8d | FP16 | [✅](models/cv/classification/resnext101_32x8d/igie/README.md#fp16) | | +| SEResNet50 | FP16 | [✅](models/cv/classification/se_resnet50/igie/README.md#fp16) | | +| ShuffleNetV1 | FP16 | | [✅](models/cv/classification/shufflenet_v1/ixrt/README.md#fp16) | +| ShuffleNetV2_x0_5 | FP16 | [✅](models/cv/classification/shufflenetv2_x0_5/igie/README.md#fp16) | | +| ShuffleNetV2_x1_0 | FP16 | [✅](models/cv/classification/shufflenetv2_x1_0/igie/README.md#fp16) | | +| ShuffleNetV2_x1_5 | FP16 | [✅](models/cv/classification/shufflenetv2_x1_5/igie/README.md#fp16) | | +| ShuffleNetV2_x2_0 | FP16 | [✅](models/cv/classification/shufflenetv2_x2_0/igie/README.md#fp16) | | +| SqueezeNet 1.0 | FP16 | [✅](models/cv/classification/squeezenet_v1_0/igie/README.md#fp16) | [✅](models/cv/classification/squeezenet_v1_0/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/squeezenet_v1_0/ixrt/README.md#int8) | +| SqueezeNet 1.1 | FP16 | | [✅](models/cv/classification/squeezenet_v1_1/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/classification/squeezenet_v1_1/ixrt/README.md#int8) | +| SVT Base | FP16 | [✅](models/cv/classification/svt_base/igie/README.md#fp16) | | +| Swin Transformer | FP16 | [✅](models/cv/classification/swin_transformer/igie/README.md#fp16) | | +| Swin Transformer Large | FP16 | | [✅](models/cv/classification/swin_transformer_large/ixrt/README.md) | +| VGG11 | FP16 | [✅](models/cv/classification/vgg11/igie/README.md#fp16) | | +| VGG16 | FP16 | [✅](models/cv/classification/vgg16/igie/README.md#fp16) | [✅](models/cv/classification/vgg16/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/vgg16/igie/README.md#int8) | | +| Wide ResNet50 | FP16 | [✅](models/cv/classification/wide_resnet50/igie/README.md#fp16) | [✅](models/cv/classification/wide_resnet50/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/classification/wide_resnet50/igie/README.md#int8) | [✅](models/cv/classification/wide_resnet50/ixrt/README.md#int8) | +| Wide ResNet101 | FP16 | [✅](models/cv/classification/wide_resnet101/igie/README.md#fp16) | | ### Detection - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
ATSSFP16Supported-
INT8--
CenterNetFP16SupportedSupported
INT8--
DETRFP16-Supported
INT8--
FCOSFP16SupportedSupported
INT8--
FoveaBoxFP16SupportedSupported
INT8--
FSAFFP16SupportedSupported
INT8--
HRNetFP16SupportedSupported
INT8--
PAAFP16Supported-
INT8--
RetinaFaceFP16SupportedSupported
INT8--
RetinaNetFP16Supported-
INT8--
RTMDetFP16Supported-
INT8--
SABLFP16Supported-
INT8--
YOLOv3FP16SupportedSupported
INT8SupportedSupported
YOLOv4FP16SupportedSupported
INT8SupportedSupported
YOLOv5FP16SupportedSupported
INT8SupportedSupported
YOLOv5sFP16-Supported
INT8-Supported
YOLOv6FP16SupportedSupported
INT8-Supported
YOLOv7FP16SupportedSupported
INT8SupportedSupported
YOLOv8FP16SupportedSupported
INT8SupportedSupported
YOLOv9FP16Supported-
INT8--
YOLOv10FP16Supported-
INT8--
YOLOv11FP16Supported-
INT8--
YOLOXFP16SupportedSupported
INT8SupportedSupported
+| Model | Prec. | IGIE | IxRT | +|------------|-------|---------------------------------------------------------|---------------------------------------------------------| +| ATSS | FP16 | [✅](models/cv/detection/atss/igie/README.md#fp16) | | +| CenterNet | FP16 | [✅](models/cv/detection/centernet/igie/README.md#fp16) | [✅](models/cv/detection/centernet/ixrt/README.md#fp16) | +| DETR | FP16 | | [✅](models/cv/detection/detr/ixrt/README.md#fp16) | +| FCOS | FP16 | [✅](models/cv/detection/fcos/igie/README.md#fp16) | [✅](models/cv/detection/fcos/ixrt/README.md#fp16) | +| FoveaBox | FP16 | [✅](models/cv/detection/foveabox/igie/README.md#fp16) | [✅](models/cv/detection/foveabox/ixrt/README.md#fp16) | +| FSAF | FP16 | [✅](models/cv/detection/fsaf/igie/README.md#fp16) | [✅](models/cv/detection/fsaf/ixrt/README.md#fp16) | +| HRNet | FP16 | [✅](models/cv/detection/hrnet/igie/README.md#fp16) | [✅](models/cv/detection/hrnet/ixrt/README.md#fp16) | +| PAA | FP16 | [✅](models/cv/detection/paa/igie/README.md#fp16) | | +| RetinaFace | FP16 | [✅](models/cv/detection/retinaface/igie/README.md#fp16) | [✅](models/cv/detection/retinaface/ixrt/README.md#fp16) | +| RetinaNet | FP16 | [✅](models/cv/detection/retinanet/igie/README.md#fp16) | | +| RTMDet | FP16 | [✅](models/cv/detection/rtmdet/igie/README.md#fp16) | | +| SABL | FP16 | [✅](models/cv/detection/sabl/igie/README.md#fp16) | | +| YOLOv3 | FP16 | [✅](models/cv/detection/yolov3/igie/README.md#fp16) | [✅](models/cv/detection/yolov3/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolov3/igie/README.md#int8) | [✅](models/cv/detection/yolov3/ixrt/README.md#int8) | +| YOLOv4 | FP16 | [✅](models/cv/detection/yolov4/igie/README.md#fp16) | [✅](models/cv/detection/yolov4/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolov4/igie/README.md#int816) | [✅](models/cv/detection/yolov4/ixrt/README.md#int816) | +| YOLOv5 | FP16 | [✅](models/cv/detection/yolov5/igie/README.md#fp16) | [✅](models/cv/detection/yolov5/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolov5/igie/README.md#int8) | [✅](models/cv/detection/yolov5/ixrt/README.md#int8) | +| YOLOv5s | FP16 | | [✅](models/cv/detection/yolov5s/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/detection/yolov5s/ixrt/README.md#int8) | +| YOLOv6 | FP16 | [✅](models/cv/detection/yolov6/igie/README.md#fp16) | [✅](models/cv/detection/yolov6/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/detection/yolov6/ixrt/README.md#int8) | +| YOLOv7 | FP16 | [✅](models/cv/detection/yolov7/igie/README.md#fp16) | [✅](models/cv/detection/yolov7/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolov7/igie/README.md#int8) | [✅](models/cv/detection/yolov7/ixrt/README.md#int8) | +| YOLOv8 | FP16 | [✅](models/cv/detection/yolov8/igie/README.md#fp16) | [✅](models/cv/detection/yolov8/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolov8/igie/README.md#int8) | [✅](models/cv/detection/yolov8/ixrt/README.md#int8) | +| YOLOv9 | FP16 | [✅](models/cv/detection/yolov9/igie/README.md#fp16) | | +| YOLOv10 | FP16 | [✅](models/cv/detection/yolov10/igie/README.md#fp16) | | +| YOLOv11 | FP16 | [✅](models/cv/detection/yolov11/igie/README.md#fp16) | | +| YOLOX | FP16 | [✅](models/cv/detection/yolox/igie/README.md#fp16) | [✅](models/cv/detection/yolox/ixrt/README.md#fp16) | +| | INT8 | [✅](models/cv/detection/yolox/igie/README.md#int8) | [✅](models/cv/detection/yolox/ixrt/README.md#int8) | ### Face Recognition - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
FaceNet -FP16-Supported
INT8-Supported
+| Model | Prec. | IGIE | IxRT | +|---------|-------|------|-------------------------------------------------| +| FaceNet | FP16 | | [✅](models/cv/face/facenet/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/cv/face/facenet/ixrt/README.md#int8) | ### OCR - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIE
Kie_layoutXLMFP16Supported
INT8-
SVTRFP16Supported
INT8-
+| Model | Prec. | IGIE | +|---------------|-----------|------------------------------------------------------| +| Kie_layoutXLM | FP16 | [✅](models/cv/ocr/kie_layoutxlm/igie/README.md#fp16) | +| SVTR | FP16 | [✅](models/cv/ocr/svtr/igie/README.md#fp16) | ### Pose Estimation - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
HRNetPoseFP16Supported-
INT8--
Lightweight OpenPoseFP16-Supported
INT8--
RTMPoseFP16SupportedSupported
INT8--
+| Model | Prec. | IGIE | IxRT | +|----------------------|-------|--------------------------------------------------------------|-------------------------------------------------------------------------| +| HRNetPose | FP16 | [✅](models/cv/pose_estimation/hrnetpose/igie/README.md#fp16) | | +| Lightweight OpenPose | FP16 | | [✅](models/cv/pose_estimation/lightweight_openpose/ixrt/README.md#fp16) | +| RTMPose | FP16 | [✅](models/cv/pose_estimation/rtmpose/igie/README.md#fp16) | [✅](models/cv/pose_estimation/rtmpose/ixrt/README.md#fp16) | ### Recommendation Systems - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
Wide & DeepFP16-Supported
INT8--
+| Model | Prec. | IGIE | IxRT | +|-------------|-------|------|------------------------------------------------------------------------| +| Wide & Deep | FP16 | | [✅](models/recommendation/ctr-prediction/widedeep/ixrt/README.md#fp16) | ### Segmentation - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
Mask R-CNNFP16-Supported
INT8--
SOLOv1FP16-Supported
INT8--
+| Model | Prec. | IGIE | IxRT | +|------------|-------|------|-----------------------------------------------------------| +| Mask R-CNN | FP16 | | [✅](models/cv/segmentation/mask_rcnn/ixrt/README.md#fp16) | +| SOLOv1 | FP16 | | [✅](models/cv/segmentation/solov1/ixrt/README.md#fp16) | ### Trace - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
FastReIDFP16Supported-
INT8--
DeepSortFP16Supported-
INT8Supported-
RepNet-Vehicle-ReIDFP16Supported-
INT8--
+| Model | Prec. | IGIE | IxRT | +|---------------------|-------|---------------------------------------------------|------| +| FastReID | FP16 | [✅](models/cv/trace/fastreid/igie/README.md#fp16) | | +| DeepSort | FP16 | [✅](models/cv/trace/deepsort/igie/README.md#fp16) | | +| | INT8 | [✅](models/cv/trace/deepsort/igie/README.md#int8) | | +| RepNet-Vehicle-ReID | FP16 | [✅](models/cv/trace/repnet/igie/README.md#fp16) | | ## LLM (Large Language Model) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsvLLMTensorRT-LLMTGI
Baichuan2-7BSupported--
ChatGLM-3-6BSupported--
ChatGLM-3-6B-32KSupported--
Llama2-7BSupportedSupported-
Llama2-13B-Supported-
Llama2-70B-Supported-
Llama3-70BSupported--
Qwen-7BSupported--
Qwen1.5-7BSupported-Supported
Qwen1.5-14BSupported--
Qwen1.5-32B ChatSupported--
Qwen1.5-72BSupported--
Qwen2-7B InstructSupported--
Qwen2-72B InstructSupported--
StableLM2-1.6BSupported--
+| Model | vLLM | TRT-LLM | TGI | +|--------------------|---------------------------------------------------------------------|------------------------------------------------------------------|-------------------------------------------------------------------------------------| +| Baichuan2-7B | [✅](models/nlp/large_language_model/baichuan2-7b/vllm/README.md) | | | +| ChatGLM-3-6B | [✅](models/nlp/large_language_model/chatglm3-6b/vllm/README.md) | | | +| ChatGLM-3-6B-32K | [✅](models/nlp/large_language_model/chatglm3-6b-32k/vllm/README.md) | | | +| Llama2-7B | [✅](models/nlp/large_language_model/llama2-7b/vllm/README.md) | [✅](models/nlp/large_language_model/llama2-7b/trtllm/README.md) | | +| Llama2-13B | | [✅](models/nlp/large_language_model/llama2-13b/trtllm/README.md) | | +| Llama2-70B | | [✅](models/nlp/large_language_model/llama2-70b/trtllm/README.md) | | +| Llama3-70B | [✅](models/nlp/large_language_model/llama3-70b/vllm/README.md) | | | +| Qwen-7B | [✅](models/nlp/large_language_model/qwen-7b/vllm/README.md) | | | +| Qwen1.5-7B | [✅](models/nlp/large_language_model/qwen1.5-7b/vllm/README.md) | | [✅](models/nlp/large_language_model/qwen1.5-7b/text-generation-inference/README.md) | +| Qwen1.5-14B | [✅](models/nlp/large_language_model/qwen1.5-14b/vllm/README.md) | | | +| Qwen1.5-32B Chat | [✅](models/nlp/large_language_model/qwen1.5-32b/vllm/README.md) | | | +| Qwen1.5-72B | [✅](models/nlp/large_language_model/qwen1.5-72b/vllm/README.md) | | | +| Qwen2-7B Instruct | [✅](models/nlp/large_language_model/qwen2-7b/vllm/README.md) | | | +| Qwen2-72B Instruct | [✅](models/nlp/large_language_model/qwen2-72b/vllm/README.md) | | | +| StableLM2-1.6B | [✅](models/nlp/large_language_model/stablelm/vllm/README.md) | | | ## Multimodal - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsvLLM
Chameleon-7BSupported
Fuyu-8BSupported
InternVL2-4BSupported
LLaVASupported
LLaVA-Next-Video-7BSupported
MiniCPM V2Supported
+| Model | vLLM | +|---------------------|-----------------------------------------------------------------------------------------| +| Chameleon-7B | [✅](models/multimodal/vision_language_understanding/chameleon_7b/vllm/README.md) | +| Fuyu-8B | [✅](models/multimodal/vision_language_understanding/fuyu_8b/vllm/README.md) | +| InternVL2-4B | [✅](models/multimodal/vision_language_understanding/intern_vl/vllm/README.md) | +| LLaVA | [✅](models/multimodal/vision_language_understanding/llava/vllm/README.md) | +| LLaVA-Next-Video-7B | [✅](models/multimodal/vision_language_understanding/llava_next_video_7b/vllm/README.md) | +| MiniCPM V2 | [✅](models/multimodal/vision_language_understanding/minicpm_v_2/vllm/README.md) | ## NLP ### Language Modelling - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
ALBERTFP16-Supported
INT8--
BERT Base NERFP16--
INT8Supported-
BERT Base SQuADFP16SupportedSupported
INT8-Supported
BERT Large SQuADFP16SupportedSupported
INT8SupportedSupported
DeBERTaFP16-Supported
INT8--
RoBERTaFP16-Supported
INT8--
RoFormerFP16-Supported
INT8--
VideoBERTFP16-Supported
INT8--
+| Model | Prec. | IGIE | IxRT | +|------------------|-------|---------------------------------------------------------------------|---------------------------------------------------------------------| +| ALBERT | FP16 | | [✅](models/nlp/language_model/albert/ixrt/README.md) | +| BERT Base NER | INT8 | [✅](models/nlp/language_model/bert_base_ner/igie/README.md#int8) | | +| BERT Base SQuAD | FP16 | [✅](models/nlp/language_model/bert_base_squad/igie/README.md#fp16) | [✅](models/nlp/language_model/bert_base_squad/ixrt/README.md#fp16) | +| | INT8 | | [✅](models/nlp/language_model/bert_base_squad/ixrt/README.md#int8) | +| BERT Large SQuAD | FP16 | [✅](models/nlp/language_model/bert_large_squad/igie/README.md#fp16) | [✅](models/nlp/language_model/bert_large_squad/ixrt/README.md#fp16) | +| | INT8 | [✅](models/nlp/language_model/bert_large_squad/igie/README.md#int8) | [✅](models/nlp/language_model/bert_large_squad/ixrt/README.md#int8) | +| DeBERTa | FP16 | | [✅](models/nlp/language_model/deberta/ixrt/README.md) | +| RoBERTa | FP16 | | [✅](models/nlp/language_model/roberta/ixrt/README.md) | +| RoFormer | FP16 | | [✅](models/nlp/language_model/roformer/ixrt/README.md) | +| VideoBERT | FP16 | | [✅](models/nlp/language_model/videobert/ixrt/README.md) | ## Speech ### Speech Recognition - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ModelsPrecisionIGIEIxRT
ConformerFP16SupportedSupported
INT8--
Transformer ASRFP16-Supported
INT8--
+| Model | Prec. | IGIE | IxRT | +|-----------------|-------|---------------------------------------------------------------------|----------------------------------------------------------------------| +| Conformer | FP16 | [✅](models/speech/speech_recognition/conformer/igie/README.md#fp16) | [✅](models/speech/speech_recognition/conformer/ixrt/README.md#fp16) | +| Transformer ASR | FP16 | | [✅](models/speech/speech_recognition/transformer_asr/ixrt/README.md) | --- diff --git a/models/multimodal/text_and_image/clip/igie/README.md b/models/cv/classification/clip/igie/README.md similarity index 100% rename from models/multimodal/text_and_image/clip/igie/README.md rename to models/cv/classification/clip/igie/README.md diff --git a/models/multimodal/text_and_image/clip/igie/build_engine.py b/models/cv/classification/clip/igie/build_engine.py similarity index 100% rename from models/multimodal/text_and_image/clip/igie/build_engine.py rename to models/cv/classification/clip/igie/build_engine.py diff --git a/models/multimodal/text_and_image/clip/igie/ci/prepare.sh b/models/cv/classification/clip/igie/ci/prepare.sh similarity index 100% rename from models/multimodal/text_and_image/clip/igie/ci/prepare.sh rename to models/cv/classification/clip/igie/ci/prepare.sh diff --git a/models/multimodal/text_and_image/clip/igie/export.py b/models/cv/classification/clip/igie/export.py similarity index 100% rename from models/multimodal/text_and_image/clip/igie/export.py rename to models/cv/classification/clip/igie/export.py diff --git a/models/multimodal/text_and_image/clip/igie/imagenet_labels.yaml b/models/cv/classification/clip/igie/imagenet_labels.yaml similarity index 100% rename from models/multimodal/text_and_image/clip/igie/imagenet_labels.yaml rename to models/cv/classification/clip/igie/imagenet_labels.yaml diff --git a/models/multimodal/text_and_image/clip/igie/inference.py b/models/cv/classification/clip/igie/inference.py similarity index 100% rename from models/multimodal/text_and_image/clip/igie/inference.py rename to models/cv/classification/clip/igie/inference.py diff --git a/models/multimodal/text_and_image/clip/igie/requirements.txt b/models/cv/classification/clip/igie/requirements.txt similarity index 100% rename from models/multimodal/text_and_image/clip/igie/requirements.txt rename to models/cv/classification/clip/igie/requirements.txt diff --git a/models/multimodal/text_and_image/clip/igie/scripts/infer_clip_fp16_accuracy.sh b/models/cv/classification/clip/igie/scripts/infer_clip_fp16_accuracy.sh similarity index 100% rename from models/multimodal/text_and_image/clip/igie/scripts/infer_clip_fp16_accuracy.sh rename to models/cv/classification/clip/igie/scripts/infer_clip_fp16_accuracy.sh diff --git a/models/multimodal/text_and_image/clip/igie/scripts/infer_clip_fp16_performance.sh b/models/cv/classification/clip/igie/scripts/infer_clip_fp16_performance.sh similarity index 100% rename from models/multimodal/text_and_image/clip/igie/scripts/infer_clip_fp16_performance.sh rename to models/cv/classification/clip/igie/scripts/infer_clip_fp16_performance.sh diff --git a/models/multimodal/text_and_image/clip/ixrt/README.md b/models/cv/classification/clip/ixformer/README.md similarity index 100% rename from models/multimodal/text_and_image/clip/ixrt/README.md rename to models/cv/classification/clip/ixformer/README.md diff --git a/models/multimodal/text_and_image/clip/ixrt/ci/prepare.sh b/models/cv/classification/clip/ixformer/ci/prepare.sh similarity index 100% rename from models/multimodal/text_and_image/clip/ixrt/ci/prepare.sh rename to models/cv/classification/clip/ixformer/ci/prepare.sh diff --git a/models/multimodal/text_and_image/clip/ixrt/inference.py b/models/cv/classification/clip/ixformer/inference.py similarity index 100% rename from models/multimodal/text_and_image/clip/ixrt/inference.py rename to models/cv/classification/clip/ixformer/inference.py diff --git a/models/nlp/large_language_model/baichuan2-7b/vllm/README.md b/models/nlp/large_language_model/baichuan2-7b/vllm/README.md index 2524abb9710b032199ee9017d4bcde4116c5cdf4..dab6b79d4b1791d3b5f92969fe2ba56be2348ba9 100755 --- a/models/nlp/large_language_model/baichuan2-7b/vllm/README.md +++ b/models/nlp/large_language_model/baichuan2-7b/vllm/README.md @@ -1,4 +1,4 @@ -# Baichuan-2-7B +# Baichuan-2-7B (vLLM) ## Description diff --git a/models/nlp/large_language_model/chatglm3-6b-32k/vllm/README.md b/models/nlp/large_language_model/chatglm3-6b-32k/vllm/README.md index bbcf0a5092a92d2001fd5a7df9775cba88538ddf..57f754042d389808bea06e2fae0a06f7c7275806 100644 --- a/models/nlp/large_language_model/chatglm3-6b-32k/vllm/README.md +++ b/models/nlp/large_language_model/chatglm3-6b-32k/vllm/README.md @@ -1,4 +1,4 @@ -# ChatGLM3-6B-32K +# ChatGLM3-6B-32K (vLLM) ## Description diff --git a/models/nlp/large_language_model/chatglm3-6b/vllm/README.md b/models/nlp/large_language_model/chatglm3-6b/vllm/README.md index 24a59ad4662d9b0000de504229c61df91c9f20da..fb782cee26b31f7c680fd8328bd4040fe9f710f1 100644 --- a/models/nlp/large_language_model/chatglm3-6b/vllm/README.md +++ b/models/nlp/large_language_model/chatglm3-6b/vllm/README.md @@ -1,4 +1,4 @@ -# ChatGLM3-6B +# ChatGLM3-6B (vLLM) ## Description diff --git a/models/nlp/large_language_model/llama2-13b/trtllm/README.md b/models/nlp/large_language_model/llama2-13b/trtllm/README.md index b24c29ce921f2571833a50d01ad3035ec28e1ea5..3c36a98f50fb440e0287e9aa889014a049153da0 100755 --- a/models/nlp/large_language_model/llama2-13b/trtllm/README.md +++ b/models/nlp/large_language_model/llama2-13b/trtllm/README.md @@ -1,4 +1,4 @@ -# Llama2 13B +# Llama2 13B (TensorRT-LLM) ## Description diff --git a/models/nlp/large_language_model/llama2-70b/trtllm/README.md b/models/nlp/large_language_model/llama2-70b/trtllm/README.md index 621dbffeeaac3387030e326e6e24f918644c9ae1..428dc7c5088b475692f7928d5b71c993f1da9f2e 100644 --- a/models/nlp/large_language_model/llama2-70b/trtllm/README.md +++ b/models/nlp/large_language_model/llama2-70b/trtllm/README.md @@ -1,4 +1,4 @@ -# LlaMa2 70B +# LlaMa2 70B (TensorRT-LLM) ## Description diff --git a/models/nlp/large_language_model/llama2-7b/trtllm/README.md b/models/nlp/large_language_model/llama2-7b/trtllm/README.md index 5e8ac1904b74f2b341d8559b41e70f0a2122a39a..bcf06d484c67b8da6e663603eed0377989e58b1a 100644 --- a/models/nlp/large_language_model/llama2-7b/trtllm/README.md +++ b/models/nlp/large_language_model/llama2-7b/trtllm/README.md @@ -1,4 +1,4 @@ -# LlaMa2 7B +# LlaMa2 7B (TensorRT-LLM) ## Description diff --git a/models/nlp/large_language_model/llama2-7b/vllm/README.md b/models/nlp/large_language_model/llama2-7b/vllm/README.md index 327936835d97bc608d58828c584dc8186e7f79ad..5fa7012733feb811fcef8992355047a5bf7a14ee 100755 --- a/models/nlp/large_language_model/llama2-7b/vllm/README.md +++ b/models/nlp/large_language_model/llama2-7b/vllm/README.md @@ -1,4 +1,4 @@ -# LlaMa2 7B +# LlaMa2 7B (vLLM) ## Description diff --git a/models/nlp/large_language_model/llama3-70b/vllm/README.md b/models/nlp/large_language_model/llama3-70b/vllm/README.md index 7df5a637cee3cdf7fd072c35a90d8cd64763b69b..8374dd94c6257e0c3e3732f3d543a97a192b2140 100644 --- a/models/nlp/large_language_model/llama3-70b/vllm/README.md +++ b/models/nlp/large_language_model/llama3-70b/vllm/README.md @@ -1,4 +1,4 @@ -# LlaMa3 70B +# LlaMa3 70B (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen-7b/vllm/README.md b/models/nlp/large_language_model/qwen-7b/vllm/README.md index 9dcb18f261d7772316fcd90e18dd5282d64d6f81..50c7cb095fc5b05b4efbad615bb7cd48652f47c8 100644 --- a/models/nlp/large_language_model/qwen-7b/vllm/README.md +++ b/models/nlp/large_language_model/qwen-7b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen-7B +# Qwen-7B (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen1.5-14b/vllm/README.md b/models/nlp/large_language_model/qwen1.5-14b/vllm/README.md index 905967c2c7372eba6bfb3791fe6833400ba68ee6..d1792860528742d77f82b9a2f68a068038d78eb2 100644 --- a/models/nlp/large_language_model/qwen1.5-14b/vllm/README.md +++ b/models/nlp/large_language_model/qwen1.5-14b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen1.5-14B +# Qwen1.5-14B (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen1.5-32b/vllm/README.md b/models/nlp/large_language_model/qwen1.5-32b/vllm/README.md index 9e48493448c984bc6798e53e150c3f95040b0a1b..4e92d517cadcd16e4943e9d3689bc1d183fd8d17 100755 --- a/models/nlp/large_language_model/qwen1.5-32b/vllm/README.md +++ b/models/nlp/large_language_model/qwen1.5-32b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen1.5-32B-Chat +# Qwen1.5-32B-Chat (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen1.5-72b/vllm/README.md b/models/nlp/large_language_model/qwen1.5-72b/vllm/README.md index 3b7c9298e78a320df17d2c1cda1fa7826c4cd73e..c9785998c8b8a196e59cb0df258a5aebe047e3d9 100644 --- a/models/nlp/large_language_model/qwen1.5-72b/vllm/README.md +++ b/models/nlp/large_language_model/qwen1.5-72b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen1.5-72B +# Qwen1.5-72B (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen1.5-7b/text-generation-inference/README.md b/models/nlp/large_language_model/qwen1.5-7b/text-generation-inference/README.md index c7dbec71a99d110e6fec603a201ecf26f7d5fcb0..41d34f2052f598deddb9cccfa41cdd367e9fad73 100644 --- a/models/nlp/large_language_model/qwen1.5-7b/text-generation-inference/README.md +++ b/models/nlp/large_language_model/qwen1.5-7b/text-generation-inference/README.md @@ -1,4 +1,4 @@ -# Qwen1.5-7B +# Qwen1.5-7B (TGI) ## Description diff --git a/models/nlp/large_language_model/qwen1.5-7b/vllm/README.md b/models/nlp/large_language_model/qwen1.5-7b/vllm/README.md index e01def64bc04d2792c5d146dac7d48e971dcfb5a..44f0c6b56fc30462118ad5862d023287c1d9ec65 100644 --- a/models/nlp/large_language_model/qwen1.5-7b/vllm/README.md +++ b/models/nlp/large_language_model/qwen1.5-7b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen1.5-7B +# Qwen1.5-7B (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen2-72b/vllm/README.md b/models/nlp/large_language_model/qwen2-72b/vllm/README.md index f925f812b5b9bd5597f9ec439e33944f419d6322..b859879019267a3cbee6eb90fae5d78c0b6fe306 100755 --- a/models/nlp/large_language_model/qwen2-72b/vllm/README.md +++ b/models/nlp/large_language_model/qwen2-72b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen2-72B-Instruct +# Qwen2-72B-Instruct (vLLM) ## Description diff --git a/models/nlp/large_language_model/qwen2-7b/vllm/README.md b/models/nlp/large_language_model/qwen2-7b/vllm/README.md index 6cfc785858a7cc6f96699f267f9b3b357ed43840..dce3c564b8c201ebee24149b0ca98b54e0bbd3b1 100755 --- a/models/nlp/large_language_model/qwen2-7b/vllm/README.md +++ b/models/nlp/large_language_model/qwen2-7b/vllm/README.md @@ -1,4 +1,4 @@ -# Qwen2-7B Instruct +# Qwen2-7B Instruct (vLLM) ## Description diff --git a/models/nlp/large_language_model/stablelm/vllm/README.md b/models/nlp/large_language_model/stablelm/vllm/README.md index e9d1b7e4fb0b7c65c7b793bf6f80cf36b667bf7e..18fd9a8f6a027e78861a0fa7ce31165e9ba019d2 100644 --- a/models/nlp/large_language_model/stablelm/vllm/README.md +++ b/models/nlp/large_language_model/stablelm/vllm/README.md @@ -1,4 +1,4 @@ -# StableLM2-1.6B +# StableLM2-1.6B (vLLM) ## Description