OpenVINO™ 向けに検証された AI モデル

以下は、OpenVINO で動作することが検証されているモデルの一覧です。OpenVINO 対応フレームワークの他のモデルも正常に動作する可能性がありますが、テストされていないことに注意してください。

OpenVINO™ ツールキットを搭載したインテル® Core™ Ultra プロセッサーで実行される AI モデル:

モデル	フレームワーク	精度	AI PC CPU	AI PC GPU	AI PC NPU
ノイズ除去	tf	FP16
ノイズ除去	tf	FP32
GPT-2	onnx	FP16
GPT-2	onnx	FP32
Sharpen-LensBlur	tf	FP16
Sharpen-LensBlur	tf	FP32
Sharpen-LensBlur	tf	FP16-INT8
Sharpen-MotionBlur	tf	FP16
シャープとモーションブラー	tf	FP16-INT8
シャープとモーションブラー	tf	FP32
Sharpen-Sharpen	tf	FP16
Sharpen-Sharpen	tf	FP16-INT8
Sharpen-Sharpen	tf	FP32
Sphereface	caffe	FP16
Sphereface	caffe	FP32
aclnet	onnx	FP16
aclnet	onnx	FP16-INT8
aclnet	onnx	FP32
aclnet-int8	onnx	FP16-INT8
alexnet-caffe	caffe	FP16
alexnet-caffe	caffe	FP32
alexnet-onnx	onnx	FP16
alexnet-onnx	onnx	FP16-INT8
alexnet-onnx	onnx	FP32
began	onnx	FP16
began	onnx	FP32
bert-base-cased-onnx	onnx	FP16
bert-base-cased-onnx	onnx	FP16-INT8
bert-base-cased-onnx	onnx	FP32
bert-base-cased-paddle	paddle	FP16
bert-base-cased-paddle	paddle	FP16-INT8
bert-base-cased-paddle	paddle	FP32
bert-base-cased-squad2	onnx	FP16
bert-base-cased-squad2	onnx	FP32
bert-base-chinese-xnli-zh	tf	FP16
bert-base-chinese-xnli-zh	tf	FP32
bert-base-ner	onnx	FP16
bert-base-ner	onnx	FP16-INT8
bert-base-ner	onnx	FP32
bert-base-uncased-mrpc	onnx	FP16
bert-base-uncased-mrpc	onnx	FP16-INT8
bert-base-uncased-mrpc	onnx	FP32
bert-large-uncased-whole-word-masking-squad-0001	onnx	FP16
bert-large-uncased-whole-word-masking-squad-0001	onnx	FP32
bert-large-uncased-whole-word-masking-squad-int8-0001	onnx	FP16-INT8
bisenet-v2	paddle	FP16
bisenet-v2	paddle	FP32
brain-tumor-segmentation-0002	onnx	FP16
brain-tumor-segmentation-0002	onnx	FP16-INT8
colorization-siggraph	onnx	FP16
colorization-siggraph	onnx	FP32
colorization-siggraph	onnx	FP16
colorization-siggraph	onnx	FP32
colorization-v2	onnx	FP16-INT8
common-sign-language-0001	onnx	FP16
common-sign-language-0001	onnx	FP32
crnn-tf	tf	FP16
crnn-tf	tf	FP32
darknet19	tf	FP16
darknet19	tf	FP16-INT8
darknet19	tf	FP32
deeplab-v3p-resnet50-os8	paddle	FP16
deeplab-v3p-resnet50-os8	paddle	FP16-INT8
deeplab-v3p-resnet50-os8	paddle	FP32
deeplabv3	tf	FP16
deeplabv3	tf	FP16-INT8
deeplabv3	tf	FP32
densenet-121-caffe	caffe	FP16
densenet-121-caffe	caffe	FP32
densenet-121-onnx	onnx	FP16
densenet-121-onnx	onnx	FP16-INT8
densenet-121-tf	tf	FP16
densenet-121-tf	tf	FP32
densenet-121-tf2	tf2	FP16
densenet-121-tf2	tf2	FP32
detr_resnet50	onnx	FP16
detr_resnet50	onnx	FP32
dien_alibaba	tf	FP16
dien_alibaba	tf	FP32
distilbert-base-cased	onnx	FP16
distilbert-base-cased	onnx	FP32
distilbert-base-uncased	onnx	FP16
distilbert-base-uncased	onnx	FP16-INT8
distilbert-base-uncased-sst-2	onnx	FP16
distilbert-base-uncased-sst-2	onnx	FP16-INT8
distilbert-base-uncased-sst-2	onnx	FP32
dna_r9.4.1@v3	onnx	FP16
dna_r9.4.1@v3	onnx	FP32
dpn-68	onnx	FP16
dpn-68	onnx	FP16-INT8
dpn-68	onnx	FP32
east_resnet_v1_50	tf	FP16
east_resnet_v1_50	tf	FP16-INT8
east_resnet_v1_50	tf	FP32
ebgan	onnx	FP16
ebgan	onnx	FP32
edsr3-nas	onnx	FP16
edsr3-nas	onnx	FP32
edsr3_super_resolution	tf	FP16
edsr3_super_resolution	tf	FP32
efficientdet-d0-tf	tf	FP16
efficientdet-d0-tf	tf	FP16-INT8
efficientdet-d0-tf	tf	FP16
efficientdet-d0-tf	tf	FP16-INT8
efficientnet-b0-onnx	onnx	FP16
efficientnet-b0-onnx	onnx	FP16-INT8
efficientnet-b0-onnx	onnx	FP32
efficientnet-b0-tf	tf	FP16
efficientnet-b0-tf	tf	FP16-INT8
efficientnet-b0-tf	tf	FP32
efficientnet-v2-m	onnx	FP16
efficientnet-v2-m	onnx	FP16-INT8
efficientnet-v2-m	onnx	FP32
emotions-recognition-retail-0003	caffe	FP16
emotions-recognition-retail-0003	caffe	FP16-INT8
esrgan	onnx	FP16
esrgan	onnx	FP32
esrgan	onnx	FP16
f3net	onnx	FP16
f3net	onnx	FP16-INT8
f3net	onnx	FP32
face-detection-0200	onnx	FP16
face-detection-0202	onnx	FP16
face-detection-0202	onnx	FP32
face-detection-0204	onnx	FP16
face-detection-adas-0001	caffe	FP16
face-detection-adas-0001	caffe	FP32
face-detection-retail-0004	caffe	FP16
face-detection-retail-0004	caffe	FP32
face-detection-retail-0044	caffe	FP16
face-detection-retail-0044	caffe	FP32
face-recognition-mobilefacenet-arcface	mxnet	FP16
face-recognition-mobilefacenet-arcface	mxnet	FP16-INT8
face-recognition-resnet50-arcface	mxnet	FP16
face-recognition-resnet50-aws	mxnet	FP16
face-recognition-resnet50-aws	mxnet	FP32
face-reidentification-retail-0095	onnx	FP16
face-reidentification-retail-0095	onnx	FP16-INT8
faceboxes-caffe	caffe	FP16
faceboxes-caffe	caffe	FP32
faceboxes-onnx	onnx	FP16
faceboxes-onnx	onnx	FP16-INT8
faceboxes-onnx	onnx	FP32
facenet	onnx	FP16
facenet	onnx	FP16-INT8
facenet-20180408-102900	tf	FP16
facenet-20180408-102900	tf	FP32
facial-landmarks-35-adas-0002	caffe	FP16
facial-landmarks-35-adas-0002	caffe	FP16-INT8
facial-landmarks-98-detection-0001	onnx	FP16
facial-landmarks-98-detection-0001	onnx	FP32
faster-rcnn-resnet101-coco-sparse-60-0001	tf	FP16
faster-rcnn-resnet101-coco-sparse-60-0001	tf	FP32
faster_rcnn_inception_resnet_v2_atrous_coco	tf	FP16
faster_rcnn_inception_resnet_v2_atrous_coco	tf	FP32
faster_rcnn_inception_v2_coco	tf	FP16
faster_rcnn_inception_v2_coco	tf	FP32
faster_rcnn_resnet101_coco	tf	FP16
faster_rcnn_resnet101_coco	tf	FP32
faster_rcnn_resnet101_kitti	tf	FP16
faster_rcnn_resnet101_kitti	tf	FP32
faster_rcnn_resnet50_coco	tf	FP16
faster_rcnn_resnet50_coco	tf	FP32
faster_rcnn_resnet50_coco	tf	FP16-INT8
faster_rcnn_resnet50_fpn_coco	onnx	FP16
fastscnn	paddle	FP16
fastscnn	paddle	FP32
fastseg-small	onnx	FP16
fastseg-small	onnx	FP16-INT8
fastseg-small	onnx	FP32
fcrn-dp-nyu-depth-v2-tf	tf	FP16
fcrn-dp-nyu-depth-v2-tf	tf	FP16-INT8
fcrn-dp-nyu-depth-v2-tf	tf	FP32
forward-tacotron-duration-prediction	onnx	FP16
forward-tacotron-duration-prediction	onnx	FP32
fsrcnn-x4	tf	FP16
fsrcnn-x4	tf	FP32
gaze-estimation-adas-0002	onnx	FP16
googlenet-v4-caffe	caffe	FP16
googlenet-v4-caffe	caffe	FP32
googlenet-v4-tf	tf	FP16
googlenet-v4-tf	tf	FP16-INT8
googlenet-v4-tf	tf	FP32
handwritten-japanese-recognition-0001	onnx	FP16-INT8
hg-s8-b1-mpii	onnx	FP16
hg-s8-b1-mpii	onnx	FP16-INT8
higher-hrnet-w32-512	onnx	FP16
higher-hrnet-w32-512	onnx	FP32
horizontal-text-detection-0001	onnx	FP16
horizontal-text-detection-0001	onnx	FP16-INT8
horizontal-text-detection-0001	onnx	FP32
human-pose-estimation-3d-0001	onnx	FP16
human-pose-estimation-3d-0001	onnx	FP32
hybrid-cs-model-mri	tf2	FP16
hybrid-cs-model-mri	tf2	FP32
i3d-flow	tf	FP16
i3d-flow	tf	FP16-INT8
i3d-flow	tf	FP32
instance-segmentation-security-0002	onnx	FP16
instance-segmentation-security-0002	onnx	FP32
instance-segmentation-security-0010	onnx	FP16
instance-segmentation-security-0010	onnx	FP32
mask_rcnn_inception_resnet_v2_atrous_coco	tf	FP16
mask_rcnn_inception_resnet_v2_atrous_coco	tf	FP32
mask_rcnn_inception_v2_coco	tf	FP16
mask_rcnn_inception_v2_coco	tf	FP16-INT8
mask_rcnn_inception_v2_coco	tf	FP32
mask_rcnn_resnet101_atrous_coco	tf	FP16
mask_rcnn_resnet101_atrous_coco	tf	FP16-INT8
mask_rcnn_resnet50_atrous_coco	tf	FP16
mask_rcnn_resnet50_atrous_coco	tf	FP16-INT8
mask_rcnn_resnet50_atrous_coco	tf	FP32
midasnet	onnx	FP16
midasnet	onnx	FP16-INT8
midasnet	onnx	FP32
mobilebert	onnx	FP16-INT8
mobilebert	onnx	FP16-INT8
mobilenet-ssd	caffe	FP16
mobilenet-ssd	caffe	FP32
mobilenet-v2-caffe	caffe	FP16
mobilenet-v2-caffe	caffe	FP32
mobilenet-v2-onnx	onnx	FP16
mobilenet-v2-onnx	onnx	FP16-INT8
mobilenet-v2-onnx	onnx	FP32
mobilenet-v2-paddle	paddle	FP16
mobilenet-v2-paddle	paddle	FP16-INT8
mobilenet-v2-paddle	paddle	FP32
mobilenet-v2-1.0-224-tf	tf	FP16
mobilenet-v2-1.0-224-tf	tf	FP16-INT8
mobilenet-v2-1.0-224-tf	tf	FP32
mobilenet-v2-1.0-224-tf2	tf2	FP16
mobilenet-v2-1.0-224-tf2	tf2	FP16-INT8
mobilenet-v2-1.0-224-tf2	tf2	FP32
mobilenet-v3-large-1.0-224-onnx	onnx	FP16
mobilenet-v3-large-1.0-224-onnx	onnx	FP16-INT8
mobilenet-v3-large-1.0-224-onnx	onnx	FP32
mobilenet-v3-large-1.0-224-paddle	paddle	FP16
mobilenet-v3-large-1.0-224-paddle	paddle	FP16-INT8
mobilenet-v3-large-1.0-224-paddle	paddle	FP32
mobilenet-v3-large-1.0-224-tf	tf	FP16
mobilenet-v3-large-1.0-224-tf	tf	FP32
mobilenet-v3-large-1.0-224-tf2	tf2	FP16
mobilenet-v3-large-1.0-224-tf2	tf2	FP16-INT8
mobilenet-v3-large-1.0-224-tf2	tf2	FP32
mobilenet-yolo-v4-syg	tf	FP16-INT8
mobilenet-yolo-v4-syg	tf	FP32
modnet_webcam_portrait_matting	onnx	FP16-INT8
mozilla-deepspeech-0.8.2	tf	FP16
mozilla-deepspeech-0.8.2	tf	FP16-INT8
mozilla-deepspeech-0.8.2	tf	FP32
ocrnet-hrnet-w18	paddle	FP16
ocrnet-hrnet-w18	paddle	FP16-INT8
ocrnet-hrnet-w18	paddle	FP32
ocrnet-hrnet-w48	paddle	FP16
ocrnet-hrnet-w48	paddle	FP32
octave-resnext-101-0.25	mxnet	FP16
octave-resnext-101-0.25	mxnet	FP32
open-closed-eye-0001	onnx	FP16
open-closed-eye-0001	onnx	FP16-INT8
open-closed-eye-0001	onnx	FP32
pedestrian-detection-adas-0002	caffe	FP16
pedestrian-detection-adas-0002	caffe	FP16-INT8
pedestrian-detection-adas-0002	caffe	FP32
person-detection-0100	onnx	FP16
person-detection-0101	onnx	FP16
person-detection-0201	onnx	FP16
person-detection-0201	onnx	FP32
person-vehicle-bike-detection-crossroad-0078	caffe	FP16
person-vehicle-bike-detection-crossroad-0078	caffe	FP16-INT8
person-vehicle-bike-detection-crossroad-0078	caffe	FP32
pp-ocr-det	paddle	FP16
pp-ocr-det	paddle	FP16-INT8
pp-ocr-det	paddle	FP32
pp-ocr-rec	paddle	FP16
pp-ocr-rec	paddle	FP32
pp-yolo	paddle	FP16
pp-yolo	paddle	FP32
prnet	tf	FP16
prnet	tf	FP16-INT8
prnet	tf	FP32
product-detection-0001	onnx	FP16
product-detection-0001	onnx	FP16-INT8
product-detection-0001	onnx	FP32
quartznet-15x5-en	onnx	FP16
quartznet-15x5-en	onnx	FP16-INT8
quartznet-15x5-en	onnx	FP32
resnet-18	onnx	FP16
resnet-18	onnx	FP16-INT8
resnet-18	onnx	FP32
resnet-50-onnx	onnx	FP16
resnet-50-onnx	onnx	FP16-INT8
resnet-50-onnx	onnx	FP32
resnet-50-paddle	paddle	FP16
resnet-50-paddle	paddle	FP16-INT8
resnet-50-paddle	paddle	FP32
resnet-50-tf	tf	FP16
resnet-50-tf	tf	FP16-INT8
resnet-50-tf	tf	FP32
resnet-50-pytorch	onnx	FP16
resnet-50-pytorch	onnx	FP16-INT8
resnet-50-pytorch	onnx	FP32
resnext-101-32x8d	onnx	FP16
resnext-101-32x8d	onnx	FP32
resnext101-32x16d-swsl	onnx	FP16
resnext101-32x16d-swsl	onnx	FP16-INT8
resnext101-32x16d-swsl	onnx	FP32
retinanet	tf2	FP16
retinanet	tf2	FP32
retinanet_resnet34	onnx	FP16
retinanet_resnet34	onnx	FP16-INT8
retinanet_resnet34	onnx	FP32
rnnt	onnx	FP16
rnnt	onnx	FP32
road-segmentation-adas-0001	caffe	FP16
roberta-base	onnx	FP16
roberta-base	onnx	FP16-INT8
roberta-base	onnx	FP32
roberta-base-mrpc	onnx	FP16-INT8
roberta-base-mrpc	onnx	FP16-INT8
roberta-large	onnx	FP16
roberta-large	onnx	FP32
sbert-base-mean-tokens	onnx	FP16
sbert-base-mean-tokens	onnx	FP16-INT8
sbert-base-mean-tokens	onnx	FP32
se-resnet-50	caffe	FP16
se-resnet-50	caffe	FP32
se-resnext-101	caffe	FP16
se-resnext-101	caffe	FP32
se-resnext-50-caffe	caffe	FP16
se-resnext-50-caffe	caffe	FP32
se-resnext-50-onnx	onnx	FP16
se-resnext-50-onnx	onnx	FP16-INT8
se-resnext-50-onnx	onnx	FP32
se-resnext-50-tf	tf	FP16
se-resnext-50-tf	tf	FP32
semantic-segmentation-adas-0001	caffe	FP16
semantic-segmentation-adas-0001	caffe	FP16-INT8
semantic-segmentation-adas-0001	caffe	FP32
shufflenet-v2-x1.0-onnx	onnx	FP16
shufflenet-v2-x1.0-onnx	onnx	FP16-INT8
shufflenet-v2-x1.0-onnx	onnx	FP32
shufflenet-v2-x1.0-tf	tf	FP16
shufflenet-v2-x1.0-tf	tf	FP16-INT8
shufflenet-v2-x1.0-tf	tf	FP32
single-image-super-resolution-1032	onnx	FP16
squeezenet1.1-caffe	caffe	FP16
squeezenet1.1-caffe	caffe	FP32
squeezenet1.1-mxnet	mxnet	FP16
squeezenet1.1-mxnet	mxnet	FP32
squeezenet1.1-onnx	onnx	FP32
srgan-onnx	onnx	FP16
srgan-tf	tf	FP16
srgan-tf	tf	FP16-INT8
srgan-tf	tf	FP32
ssd-resnet34-1200	onnx	FP16
ssd-resnet34-1200	onnx	FP16-INT8
ssd-resnet34-1200	onnx	FP32
ssd300-onnx-0001	onnx	FP16
ssd300-onnx-0001	onnx	FP16-INT8
ssd_mobilenet_v1_coco	tf	FP16
ssd_mobilenet_v1_coco	tf	FP16-INT8
ssd_mobilenet_v1_coco	tf	FP32
ssdlite-mobilenet-v3-small-320-coco	paddle	FP16
ssdlite-mobilenet-v3-small-320-coco	paddle	FP32
ssdlite_mobilenet_v2	tf	FP16
ssdlite_mobilenet_v2	tf	FP16-INT8
ssdlite_mobilenet_v2	tf	FP32
style_transfer	onnx	FP16
style_transfer	onnx	FP16-INT8
style_transfer	onnx	FP32
stylegan2	tf	FP16
stylegan2	tf	FP32
swin-tiny-patch4-window7-224	onnx	FP16
swin-tiny-patch4-window7-224	onnx	FP16-INT8
swin-tiny-patch4-window7-224	onnx	FP32
t2t-vit-7	onnx	FP16
t2t-vit-7	onnx	FP16-INT8
t2t-vit-7	onnx	FP32
tedlium_dnn4_smbr	kaldi	FP16
tedlium_dnn4_smbr	kaldi	FP16-INT8
tedlium_dnn4_smbr	kaldi	FP32
text-image-super-resolution-0001	onnx	FP16
text-image-super-resolution-0001	onnx	FP16-INT8
text-image-super-resolution-0001	onnx	FP32
text-recognition-0012	tf	FP16
text-recognition-0012	tf	FP32
text-recognition-0013	onnx	FP16
text-recognition-0013	onnx	FP32
text-recognition-0014	onnx	FP16
text-recognition-0014	onnx	FP16-INT8
text-recognition-0014	onnx	FP32
text-to-speech-en-multi-0001	null	FP32
tiny_yolo_v2	onnx	FP16
tiny_yolo_v2	onnx	FP32
tinybert_6layer_768dim_cola	onnx	FP16
tinybert_6layer_768dim_cola	onnx	FP16-INT8
tinybert_6layer_768dim_cola	onnx	FP32
tiramisu-fc-densenet-103	tf2	FP16
tiramisu-fc-densenet-103	tf2	FP16-INT8
tiramisu-fc-densenet-103	tf2	FP32
topaz_video_super_resolution	tf	FP16
topaz_video_super_resolution	tf	FP32
ultra-lightweight-face-detection-rfb-320	onnx	FP16
ultra-lightweight-face-detection-slim-320	onnx	FP16
unet-2d	onnx	FP16
unet-camvid-onnx-0001	onnx	FP16
unet-camvid-onnx-0001	onnx	FP16-INT8
unet-camvid-onnx-0001	onnx	FP32
unet3d_mlperf-onnx	onnx	FP16
unet3d_mlperf-onnx	onnx	FP16-INT8
unet3d_mlperf-onnx	onnx	FP32
unet3d_mlperf-tf	tf	FP16
unet3d_mlperf-tf	tf	FP16-INT8
urlnet	tf2	FP16
urlnet	tf2	FP32
vehicle-detection-adas-0002	caffe	FP16
vitstr-small-patch16-224	onnx	FP16
vitstr-small-patch16-224	onnx	FP32
wav2vec2-base	onnx	FP16
wav2vec2-base	onnx	FP16-INT8
wav2vec2-base	onnx	FP32
wavernn-rnn	onnx	FP16
wavernn-rnn	onnx	FP32
wavernn-upsampler	onnx	FP16
wavernn-upsampler	onnx	FP32
wdsr-small-x4	onnx	FP16
wdsr-small-x4	onnx	FP16-INT8
wdsr-small-x4	onnx	FP32
weld-porosity-detection-0001	onnx	FP16
weld-porosity-detection-0001	onnx	FP32
xlm-roberta-base	onnx	FP16
xlm-roberta-base	onnx	FP32
yolo-v2-ava-0001	tf	FP16
yolo-v2-ava-0001	tf	FP16-INT8
yolo-v2-ava-0001	tf	FP32
yolo-v2-ava-sparse-70-0001	tf	FP16
yolo-v2-ava-sparse-70-0001	tf	FP16-INT8
yolo-v2-ava-sparse-70-0001	tf	FP32
yolo-v3-caffe	caffe	FP16
yolo-v3-caffe	caffe	FP32
yolo-v3-tf	tf	FP16
yolo-v3-tf	tf	FP32
yolo_v3-onnx	onnx	FP16
yolo_v3-onnx	onnx	FP16
yolo_v3_tiny	onnx	FP16
yolo_v3_tiny-onnx	onnx	FP16
yolo_v3_tiny-onnx	onnx	FP16-INT8
yolo_v3_tiny-onnx	onnx	FP32
yolo_v3_tiny-tf	tf	FP16
yolo_v3_tiny-tf	tf	FP16-INT8
yolo_v3_tiny-tf	tf	FP32
yolo_v4-tf	tf	FP16
yolo_v4-tf	tf	FP16-INT8
yolo_v4-tf	tf	FP32
yolo_v4-tf2	tf2	FP16
yolo_v4-tf2	tf2	FP16-INT8
yolo_v5l	onnx	FP16
yolo_v5l	onnx	FP16-INT8
yolo_v5l	onnx	FP32
yolo_v5m	onnx	FP16
yolo_v5m	onnx	FP16-INT8
yolo_v5m	onnx	FP32
yolo_v5m-v6	onnx	FP16
yolo_v5m-v6	onnx	FP16-INT8
yolo_v5m-v6	onnx	FP32
yolo_v5s	onnx	FP16
yolo_v5s	onnx	FP16-INT8
yolo_v5s	onnx	FP32
yolo_v5s-v6	onnx	FP16
yolo_v5s-v6	onnx	FP16-INT8
yolo_v5s-v6	onnx	FP32
yolo_v5x	onnx	FP16
yolo_v5x	onnx	FP32
yolo_v8n	onnx	FP16
yolo_v8n	onnx	FP16-INT8
yolo_v8n	onnx	FP32
yolof	onnx	FP16
yolof	onnx	FP16-INT8
yolof	onnx	FP32
yolor_p6	onnx	FP16
yolor_p6	onnx	FP32
yolor_p6	onnx	FP16-INT8
arcfaceresnet100-8_resnet100	onnx	FP32
bertsquad-12	onnx	FP32
bvlcalexnet-12	onnx	FP32
caffenet-12	onnx	FP32
candy-8_candy	onnx	FP32
efficientnet-lite4-11_efficientnet-lite4	onnx	FP32
emotion-ferplus-8_model	onnx	FP32
fcn-resnet101-11	onnx	FP32
fcn-resnet50-11	onnx	FP32
googlenet-12	onnx	FP32
gpt2-lm-head-10_GPT-2-LM-HEAD	onnx	FP32
inception-v1-12	onnx	FP32
inception-v2-7_model	onnx	FP32
MaskRCNN-12	onnx	FP32
mnist-12	onnx	FP32
mosaic-8_mosaic	onnx	FP32
pointilism-8_pointilism	onnx	FP32
rain-princess-8_rain_princess	onnx	FP32
rcnn-ilsvrc13-7_model	onnx	FP32
ResNet101-DUC-12	onnx	FP32
resnet101-v1-7	onnx	FP32
resnet101-v2-7	onnx	FP32
resnet152-v1-7	onnx	FP32
resnet152-v2-7	onnx	FP32
resnet18-v1-7	onnx	FP32
resnet18-v2-7	onnx	FP32
resnet34-v1-7	onnx	FP32
resnet34-v2-7	onnx	FP32
resnet50-caffe2-v1-7_model	onnx	FP32
resnet50-v2-7	onnx	FP32
retinanet-9_test_retinanet_resnet101	onnx	FP32
shufflenet-7_model	onnx	FP32
squeezenet1.0-12	onnx	FP32
ssd-10	onnx	FP32
super-resolution-10_super_resolution	onnx	FP32
t5-decoder-with-lm-head-12_t5-decoder-with-lm-head	onnx	FP32
t5-encoder-12_t5-encoder	onnx	FP32
udnie-8_udnie	onnx	FP32
version-RFB-320	onnx	FP32
version-RFB-640	onnx	FP32
vgg16-12	onnx	FP32
vgg19-7_vgg19	onnx	FP32
vgg19-caffe2-7_model	onnx	FP32
yolov2-coco-9	onnx	FP32
yolov4	onnx	FP32
zfnet512-12	onnx	FP32
arcfaceresnet100-8_resnet100	onnx	FP16
bertsquad-12	onnx	FP16
bvlcalexnet-12	onnx	FP16
caffenet-12	onnx	FP16
candy-8_candy	onnx	FP16
efficientnet-lite4-11_efficientnet-lite4	onnx	FP16
emotion-ferplus-8_model	onnx	FP16
fcn-resnet101-11	onnx	FP16
fcn-resnet50-11	onnx	FP16
googlenet-12	onnx	FP16
gpt2-lm-head-10_GPT-2-LM-HEAD	onnx	FP16
inception-v1-12	onnx	FP16
inception-v2-7_model	onnx	FP16
MaskRCNN-12	onnx	FP16
mnist-12	onnx	FP16
mosaic-8_mosaic	onnx	FP16
pointilism-8_pointilism	onnx	FP16
rain-princess-8_rain_princess	onnx	FP16
rcnn-ilsvrc13-7_model	onnx	FP16
ResNet101-DUC-12	onnx	FP16
resnet101-v1-7	onnx	FP16
resnet101-v2-7	onnx	FP16
resnet152-v1-7	onnx	FP16
resnet152-v2-7	onnx	FP16
resnet18-v1-7	onnx	FP16
resnet18-v2-7	onnx	FP16
resnet34-v1-7	onnx	FP16
resnet34-v2-7	onnx	FP16
resnet50-caffe2-v1-7_model	onnx	FP16
resnet50-v2-7	onnx	FP16
retinanet-9_test_retinanet_resnet101	onnx	FP16
shufflenet-7_model	onnx	FP16
squeezenet1.0-12	onnx	FP16
ssd_mobilenet_v1_10_ssd_mobilenet_v1	onnx	FP16
ssd-10	onnx	FP16
super-resolution-10_super_resolution	onnx	FP16
t5-decoder-with-lm-head-12_t5-decoder-with-lm-head	onnx	FP16
t5-encoder-12_t5-encoder	onnx	FP16
udnie-8_udnie	onnx	FP16
version-RFB-320	onnx	FP16
version-RFB-640	onnx	FP16
vgg16-12	onnx	FP16
vgg19-7_vgg19	onnx	FP16
vgg19-caffe2-7_model	onnx	FP16
yolov2-coco-9	onnx	FP16
yolov4	onnx	FP16
zfnet512-12	onnx	FP16
adv_inception_v3	pytorch	FP32
beit_base_patch16_224	pytorch	FP32
botnet26t_256	pytorch	FP32
cait_m36_384	pytorch	FP32
coat_lite_mini	pytorch	FP32
convit_base	pytorch	FP32
convmixer_768_32	pytorch	FP32
convnext_base	pytorch	FP32
crossvit_9_240	pytorch	FP32
cspdarknet53	pytorch	FP32
deit_base_distilled_patch16_224	pytorch	FP32
dla102	pytorch	FP32
dm_nfnet_f0	pytorch	FP32
dpn107	pytorch	FP32
eca_botnext26ts_256	pytorch	FP32
eca_halonext26ts	pytorch	FP32
ese_vovnet19b_dw	pytorch	FP32
fbnetc_100	pytorch	FP32
fbnetv3_b	pytorch	FP32
gernet_l	pytorch	FP32
ghostnet_100	pytorch	FP32
gluon_inception_v3	pytorch	FP32
gmixer_24_224	pytorch	FP32
gmlp_s16_224	pytorch	FP32
hrnet_w18	pytorch	FP32
inception_v3	pytorch	FP32
jx_nest_base	pytorch	FP32
lcnet_050	pytorch	FP32
levit_128	pytorch	FP32
mixer_b16_224	pytorch	FP32
mixnet_l	pytorch	FP32
mnasnet_100	pytorch	FP32
mobilenetv2_100	pytorch	FP32
mobilenetv3_large_100	pytorch	FP32
mobilevit_s	pytorch	FP32
nfnet_l0	pytorch	FP32
pit_b_224	pytorch	FP32
pnasnet5large	pytorch	FP32
poolformer_m36	pytorch	FP32
regnety_002	pytorch	FP32
repvgg_a2	pytorch	FP32
res2net101_26w_4s	pytorch	FP32
res2net50_14w_8s	pytorch	FP32
res2next50	pytorch	FP32
resmlp_12_224	pytorch	FP32
resnest101e	pytorch	FP32
rexnet_100	pytorch	FP32
sebotnet33ts_256	pytorch	FP32
selecsls42b	pytorch	FP32
spnasnet_100	pytorch	FP32
swin_base_patch4_window7_224	pytorch	FP32
swsl_resnext101_32x16d	pytorch	FP32
tf_efficientnet_b0	pytorch	FP32
tf_mixnet_l	pytorch	FP32
tinynet_a	pytorch	FP32
tnt_s_patch16_224	pytorch	FP32
twins_pcpvt_base	pytorch	FP32
visformer_small	pytorch	FP32
vit_base_patch16_224	pytorch	FP32
volo_d1_224	pytorch	FP32
xcit_large_24_p8_224	pytorch	FP32
albert-base-v2	pytorch	intel-optimum default
albert-base-v2-sst2	pytorch	intel-optimum default
albert-large-v2	pytorch	intel-optimum default
albert-large-v2-finetuned-mnli	pytorch	intel-optimum default
Aquila2-7B	pytorch	intel-optimum default
Aquila-7B	pytorch	intel-optimum default
AquilaChat2-7B	pytorch	intel-optimum default
AquilaChat-7B	pytorch	intel-optimum default
ast-finetuned-speech-commands-v2	pytorch	intel-optimum default
Baichuan2-13B-Chat	pytorch	intel-optimum default
Baichuan2-7B-Base	pytorch	intel-optimum default
Baichuan2-7B-Chat	pytorch	intel-optimum default
bart-base-japanese	pytorch	intel-optimum default
beit-base-patch16-224-pt22k-ft22k	pytorch	intel-optimum default
bert-base-cased-conversational	pytorch	intel-optimum default
bert-base-NER	pytorch	intel-optimum default
bert-base-uncased	pytorch	intel-optimum default
bert-base-uncased-yelp-polarity	pytorch	intel-optimum default
BERT-Large-CT-STSb	pytorch	intel-optimum default
bert-large-NER	pytorch	intel-optimum default
bert-large-uncased	pytorch	intel-optimum default
bert-large-uncased-whole-word-masking-finetuned-sst-2	pytorch	intel-optimum default
bert-small-finetuned-squadv2	pytorch	intel-optimum default
bge-base-en-v1.5	pytorch	intel-optimum default
bge-large-en-v1.5	pytorch	intel-optimum default
bge-reranker-base	pytorch	intel-optimum default
bge-reranker-large	pytorch	intel-optimum default
BioMedLM	pytorch	intel-optimum default
blenderbot-400M-distill	pytorch	intel-optimum default
bloom-560m	pytorch	intel-optimum default
bloomz-1b1	pytorch	intel-optimum default
bloomz-3b	pytorch	intel-optimum default
bloomz-7b1-4800MT	pytorch	intel-optimum default
codegemma-1.1-2b	pytorch	intel-optimum default
codegemma-1.1-7b-it	pytorch	intel-optimum default
codegemma-2b	pytorch	intel-optimum default
codegemma-7b	pytorch	intel-optimum default
codegen2-1B_P	pytorch	intel-optimum default
codegen2-3_7B_P	pytorch	intel-optimum default
codegen2-7B_P	pytorch	intel-optimum default
codegen-2B-multi	pytorch	intel-optimum default
codegen-350M-mono	pytorch	intel-optimum default
codegen-6B-multi	pytorch	intel-optimum default
CodeQwen1.5-7B-Chat	pytorch	intel-optimum default
convbert-base-turkish-mc4-toxicity-uncased	pytorch	intel-optimum default
convnext-base-224	pytorch	intel-optimum default
DPT-v3-Large	pytorch	intel-optimum default
distilbert-base-nli-mean-tokens	pytorch	intel-optimum default
distilbert-base-uncased-distilled-squad	pytorch	intel-optimum default
distilbert-NER	pytorch	intel-optimum default
distil-large-v3	pytorch	intel-optimum default
distil-medium.en	pytorch	intel-optimum default
electra-base-french-europeana-cased-generator	pytorch	intel-optimum default
electra-large-discriminator-nli-efl-tweeteval	pytorch	intel-optimum default
falcon-11B	pytorch	intel-optimum default
falcon-7b	pytorch	intel-optimum default
falcon-7b-instruct	pytorch	intel-optimum default
flaubert_base_cased	pytorch	intel-optimum default
flaubert-base-uncased-finetuned-cooking	pytorch	intel-optimum default
gemma-1.1-2b-it	pytorch	intel-optimum default
gemma-1.1-7b-it	pytorch	intel-optimum default
gemma-2b	pytorch	intel-optimum default
gemma-2b-it	pytorch	intel-optimum default
gemma-7b	pytorch	intel-optimum default
gemma-7b-it	pytorch	intel-optimum default
GPT2Neo1.3BPoints	pytorch	intel-optimum default
granite-3b-code-base	pytorch	intel-optimum default
granite-3b-code-instruct	pytorch	intel-optimum default
granite-8b-code-base	pytorch	intel-optimum default
granite-8b-code-instruct	pytorch	intel-optimum default
hubert-large-ls960-ft	pytorch	intel-optimum default
ibert-roberta-base	pytorch	intel-optimum default
ibert-roberta-base-finetuned-mrpc	pytorch	intel-optimum default
internlm2-1_8b	pytorch	intel-optimum default
internlm2-7b	pytorch	intel-optimum default
internlm2-chat-1_8b	pytorch	intel-optimum default
internlm2-chat-7b	pytorch	intel-optimum default
internlm2-chat-7b-sft	pytorch	intel-optimum default
internlm2-math-7b	pytorch	intel-optimum default
internlm2-math-base-7b	pytorch	intel-optimum default
klue-roberta-base-ner	pytorch	intel-optimum default
LCM_Dreamshaper_v7	pytorch	intel-optimum default
levit-256	pytorch	intel-optimum default
Llama-2-13b-chat-hf	pytorch	intel-optimum default
Llama-2-13b-hf	pytorch	intel-optimum default
Llama-2-7b-chat-hf	pytorch	intel-optimum default
Llama-2-7b-hf	pytorch	intel-optimum default
m2m100_418M	pytorch	intel-optimum default
Magicoder-CL-7B	pytorch	intel-optimum default
Magicoder-DS-6.7B	pytorch	intel-optimum default
Magicoder-S-CL-7B	pytorch	intel-optimum default
Magicoder-S-DS-6.7B	pytorch	intel-optimum default
MarianCausalLM	pytorch	intel-optimum default
marian-finetuned-kde4-en-to-cn-accelerate	pytorch	intel-optimum default
mbart-large-50-many-to-one-mmt	pytorch	intel-optimum default
Meta-Llama-3-8B	pytorch	intel-optimum default
Meta-Llama-3-8B-Instruct	pytorch	intel-optimum default
Meta-Llama-Guard-2-8B	pytorch	intel-optimum default
MindChat-Qwen2-4B	pytorch	intel-optimum default
MiniCPM-2B-sft-bf16	pytorch	intel-optimum default
Mistral-7B-Instruct-v0.2	pytorch	intel-optimum default
Mistral-7B-Instruct-v0.3	pytorch	intel-optimum default
Mistral-7B-v0.3	pytorch	intel-optimum default
mobilevit-xx-small	pytorch	intel-optimum default
mpt-7b	pytorch	intel-optimum default
mpt-7b-8k	pytorch	intel-optimum default
mpt-7b-8k-chat	pytorch	intel-optimum default
mpt-7b-chat	pytorch	intel-optimum default
mpt-7b-instruct	pytorch	intel-optimum default
mpt-7b-storywriter	pytorch	intel-optimum default
ms-marco-electra-base	pytorch	intel-optimum default
mt5-base	pytorch	intel-optimum default
mt5-large	pytorch	intel-optimum default
mt5-small	pytorch	intel-optimum default
neural-chat-7b-v1-1	pytorch	intel-optimum default
neural-chat-7b-v3-3	pytorch	intel-optimum default
openchat-3.6-8b-20240522	pytorch	intel-optimum default
opt-1.3b	pytorch	intel-optimum default
opt-125m	pytorch	intel-optimum default
opt-13b	pytorch	intel-optimum default
opt-2.7b	pytorch	intel-optimum default
opt-350m	pytorch	intel-optimum default
opt-6.7b	pytorch	intel-optimum default
opt-iml-1.3b	pytorch	intel-optimum default
Pegasus-7b-slerp	pytorch	intel-optimum default
pegasus-samsum	pytorch	intel-optimum default
persimmon-8b-chat	pytorch	intel-optimum default
phi-2	pytorch	intel-optimum default
Phi-3-medium-4k-instruct	pytorch	intel-optimum default
Phi-3-mini-128k-instruct	pytorch	intel-optimum default
pix2struct-base	pytorch	intel-optimum default
Polish_RoBERTa_large_OPI	pytorch	intel-optimum default
pythia-1.4b	pytorch	intel-optimum default
pythia-12b	pytorch	intel-optimum default
pythia-14m	pytorch	intel-optimum default
pythia-160m	pytorch	intel-optimum default
pythia-1b	pytorch	intel-optimum default
pythia-2.8b	pytorch	intel-optimum default
pythia-410m	pytorch	intel-optimum default
pythia-6.9b	pytorch	intel-optimum default
pythia-70m	pytorch	intel-optimum default
Pythia-Chat-Base-7B	pytorch	intel-optimum default
Qwen1.5-0.5B	pytorch	intel-optimum default
Qwen1.5-0.5B-Chat	pytorch	intel-optimum default
Qwen1.5-1.8B	pytorch	intel-optimum default
Qwen1.5-1.8B-Chat	pytorch	intel-optimum default
Qwen1.5-4B	pytorch	intel-optimum default
Qwen1.5-4B-Chat	pytorch	intel-optimum default
Qwen1.5-7B	pytorch	intel-optimum default
Qwen1.5-7B-Chat	pytorch	intel-optimum default
Qwen-1_8B	pytorch	intel-optimum default
Qwen-1_8B-Chat	pytorch	intel-optimum default
Qwen-7B	pytorch	intel-optimum default
Qwen-7B-Chat	pytorch	intel-optimum default
RedPajama-INCITE-7B-Base	pytorch	intel-optimum default
RedPajama-INCITE-7B-Chat	pytorch	intel-optimum default
RedPajama-INCITE-7B-Instruct	pytorch	intel-optimum default
RedPajama-INCITE-Chat-3B-v1	pytorch	intel-optimum default
roberta-base-squad2	pytorch	intel-optimum default
roberta-large-mnli	pytorch	intel-optimum default
roberta-large-ner-english	pytorch	intel-optimum default
roberta-large-squadv2	pytorch	intel-optimum default
roformer_small_generator	pytorch	intel-optimum default
stable-diffusion-2-1	pytorch	intel-optimum default
stable-diffusion-v1-4	pytorch	intel-optimum default
stable-diffusion-v1-5	pytorch	intel-optimum default
stable-diffusion-xl-base-1.0	pytorch	intel-optimum default
stablelm-2-1_6b	pytorch	intel-optimum default
stablelm-2-1_6b-chat	pytorch	intel-optimum default
stablelm-2-12b	pytorch	intel-optimum default
stablelm-2-12b-chat	pytorch	intel-optimum default
stablelm-2-zephyr-1_6b	pytorch	intel-optimum default
stablelm-3b-4e1t	pytorch	intel-optimum default
stablelm-base-alpha-3b	pytorch	intel-optimum default
stablelm-tuned-alpha-7b	pytorch	intel-optimum default
stablelm-zephyr-3b	pytorch	intel-optimum default
sup-simcse-roberta-base	pytorch	intel-optimum default
t5-base	pytorch	intel-optimum default
t5-large	pytorch	intel-optimum default
t5-small	pytorch	intel-optimum default
trocr-base-handwritten	pytorch	intel-optimum default
twitter-roberta-base-sentiment-latest	pytorch	intel-optimum default
unispeech-1350-en-90-it-ft-1h	pytorch	intel-optimum default
unispeech-sat-base-plus-sd	pytorch	intel-optimum default
unispeech-sat-base-sv	pytorch	intel-optimum default
unispeech-sat-large-sv	pytorch	intel-optimum default
wav2vec2-base-superb-sd	pytorch	intel-optimum default
wav2vec2-base-superb-sid	pytorch	intel-optimum default
wav2vec2-base-superb-sv	pytorch	intel-optimum default
whisper-base	pytorch	intel-optimum default
whisper-large-v2	pytorch	intel-optimum default
whisper-medium	pytorch	intel-optimum default
whisper-small	pytorch	intel-optimum default
WizardMath-7B-V1.1	pytorch	intel-optimum default
xlm-clm-ende-1024	pytorch	intel-optimum default
xlm-roberta-base-language-detection	pytorch	intel-optimum default
xlm-roberta-base-uncased-conll2003	pytorch	intel-optimum default
xlm-roberta-large	pytorch	intel-optimum default
xlm-roberta-large-en-ru-mnli	pytorch	intel-optimum default
xlm-roberta-large-finetuned-conll03-english	pytorch	intel-optimum default
XVERSE-7B-Chat	pytorch	intel-optimum default
zephyr-7b-beta	pytorch	intel-optimum default
tiny-random-albert	pytorch	intel-optimum default
tiny-random-bert	pytorch	intel-optimum default
tiny-random-ConvBertForSequenceClassification	pytorch	intel-optimum default
tiny-random-distilbert	pytorch	intel-optimum default
tiny-random-electra	pytorch	intel-optimum default
tiny-random-flaubert	pytorch	intel-optimum default
tiny-random-ibert	pytorch	intel-optimum default
tiny-random-roberta	pytorch	intel-optimum default
tiny-random-roformer	pytorch	intel-optimum default
tiny-random-squeezebert	pytorch	intel-optimum default
tiny-random-xlm	pytorch	intel-optimum default
stsb-bert-tiny-safetensors	pytorch	intel-optimum default
tiny-random-bart	pytorch	INT4
tiny-random-baichuan2	pytorch	INT4
tiny-random-baichuan2-13b	pytorch	INT4
tiny-random-BlenderbotModel	pytorch	INT4
tiny-random-BloomModel	pytorch	INT4
tiny-random-chatglm2	pytorch	INT4
tiny-random-CodeGenForCausalLM	pytorch	INT4
tiny-random-codegen2	pytorch	INT4
tiny-random-GemmaForCausalLM	pytorch	INT4
tiny-random-gpt2	pytorch	INT4
tiny-random-GPTNeoModel	pytorch	INT4
tiny-random-GPTNeoXForCausalLM	pytorch	INT4
tiny-llama-fast-tokenizer	pytorch	INT4
tiny-marian-en-de	pytorch	INT4
tiny-random-minicpm	pytorch	INT4
tiny-random-mistral	pytorch	INT4
tiny-mixtral	pytorch	INT4
tiny-random-MptForCausalLM	pytorch	INT4
tiny-random-olmo-hf	pytorch	INT4
tiny-random-OPTModel	pytorch	INT4
tiny-random-pegasus	pytorch	INT4
tiny-random-qwen	pytorch	INT4
tiny-random-StableLmForCausalLM	pytorch	INT4
tiny-random-Starcoder2ForCausalLM	pytorch	INT4
tiny-random-PhiForCausalLM	pytorch	INT4
tiny-random-phi3	pytorch	INT4
tiny-random-internlm2	pytorch	INT4
tiny-random-orion	pytorch	INT4
really-tiny-falcon-testing	pytorch	INT4
tiny-random-falcon-40b	pytorch	INT4
tiny-random-PersimmonForCausalLM	pytorch	INT4
tiny-random-CohereForCausalLM	pytorch	INT4
tiny-random-aquilachat	pytorch	INT4
tiny-random-aquila2	pytorch	INT4
tiny-random-xverse	pytorch	INT4
tiny-random-internlm	pytorch	INT4
tiny-random-dbrx	pytorch	INT4
tiny-random-qwen1.5-moe	pytorch	INT4
tiny-random-deberta	pytorch	intel-optimum default
tiny-xlm-roberta	pytorch	intel-optimum default
tiny-random-BeitForImageClassification	pytorch	intel-optimum default
tiny-random-convnext	pytorch	intel-optimum default
tiny-random-LevitModel	pytorch	intel-optimum default
tiny-random-MobileNetV2Model	pytorch	intel-optimum default
tiny-random-mobilevit	pytorch	intel-optimum default
tiny-random-resnet	pytorch	intel-optimum default
tiny-random-vit	pytorch	intel-optimum default
tiny-random-m2m_100	pytorch	intel-optimum default
tiny-random-mbart	pytorch	intel-optimum default
mt5-tiny-random	pytorch	intel-optimum default
tiny-random-t5	pytorch	intel-optimum default
tiny-random-unispeech	pytorch	intel-optimum default
wav2vec2-random-tiny-classifier	pytorch	intel-optimum default
tiny-random-Data2VecAudioModel	pytorch	intel-optimum default
tiny-random-HubertModel	pytorch	intel-optimum default
tiny-random-SEWModel	pytorch	intel-optimum default
sew-d-tiny-100k-ft-ls100h	pytorch	intel-optimum default
tiny-random-UnispeechSatModel	pytorch	intel-optimum default
tiny-random-WavlmModel	pytorch	intel-optimum default
tiny-random-Wav2Vec2Model	pytorch	intel-optimum default
tiny-random-wav2vec2-conformer	pytorch	intel-optimum default
pix2struct-tiny-random	pytorch	intel-optimum default
whisper-tiny.en	pytorch	intel-optimum default
tiny-random-VisionEncoderDecoderModel-vit-gpt2	pytorch	intel-optimum default
trocr-small-handwritten	pytorch	intel-optimum default
tiny-doc-qa-vision-encoder-decoder	pytorch	intel-optimum default
vit-with-attentions	pytorch	intel-optimum default
vit-with-hidden_states	pytorch	intel-optimum default
tiny-stable-diffusion-torch	pytorch	intel-optimum default
tiny-random-stable-diffusion-xl	pytorch	intel-optimum default
tiny-random-stable-diffusion-xl-refiner	pytorch	intel-optimum default
tiny-random-latent-consistency	pytorch	intel-optimum default

マークされたセルは、エラーなしで推論にパスしたモデルを示します。空のセルはテストされていないモデルを示します。エラーが発生し失敗した実行は記録されていません。

精度カラムの “optimum-intel default” ラベルは、小規模モデルの場合は FP32 に、1B パラメーターを超えるモデルの場合は INT8 に対応します。

2024 年 6 月 17 日時点の OpenVINO バージョン 2024.2 の結果。

モデルは、Pytorch Model Zoo や HuggingFace などのさまざまなパブリック・モデル・リポジトリーから取得され、OpenVINO を使用してネイティブまたはバックエンドとして指定されたハードウェア上で実行されました。