Model Accuracy¶
The following two tables present the absolute accuracy drop calculated as the accuracy difference between OV-accuracy and the original frame work accuracy for FP32, and the same for INT8, BF16 and FP16 representations of a model on three platform architectures. Please also refer to notes below The following two tables present the absolute accuracy drop calculated as the accuracy difference between OV-accuracy and the original frame work accuracy for FP32, and the same for INT8, BF16 and FP16 representations of a model on three platform architectures. Please also refer to notes below the table for more information.
A - Intel® Core™ i9-9000K (AVX2), INT8 and FP32
B - Intel® Xeon® 6338, (VNNI), INT8 and FP32
C - Intel(R) Xeon 8490H (VNNI, AMX), INT8, BF16, FP32
D - Intel® Flex-170, INT8 and FP16
OpenVINO™ Model name |
dataset |
Metric Name |
A, INT8 |
B, INT8 |
C, INT8 |
D, INT8 |
|---|---|---|---|---|---|---|
bert-base-cased |
SST-2_bert_cased_padded |
3.17% |
2.68% |
3.00% |
2.73% |
|
bert-large-uncased-whole-word-masking-squad-0001 |
SQUAD_v1_1_bert_msl384_mql64_ds128_lowercase |
F1 |
0.07% |
-0.03% |
0.13% |
0.11% |
efficientdet-d0 |
COCO2017_detection_91cl |
coco_precision |
-0.84% |
-0.59% |
-0.62% |
-0.63% |
mask_rcnn_resnet50_atrous_coco |
COCO2017_detection_91cl_bkgr |
coco_orig_precision |
0.03% |
0.08% |
0.11% |
0.07% |
mobilenet-v2 |
ImageNet2012 |
accuracy @ top1 |
% |
-0.97% |
-0.97% |
-0.95% |
resnet-50 |
ImageNet2012 |
accuracy @ top1 |
-0.20% |
-0.19% |
-0.13% |
-0.15% |
ssd-resnet34-1200 |
COCO2017_detection_80cl_bkgr |
map |
-0.03% |
-0.06% |
-0.01% |
0.04% |
ssd-mobilenet-v1-coco |
COCO2017_detection_80cl_bkgr |
coco-precision |
-2.97% |
-0.29% |
-0.31% |
-0.26% |
unet-camvid-onnx-0001 |
CamVid_12cl |
mean_iou @ mean |
-6.32% |
6.40% |
6.41% |
6.40% |
yolo_v3_tiny |
COCO2017_detection_80cl |
map |
% |
-0.23% |
-0.24% |
-0.66% |
yolo_v8n |
COCO2017_detection_80cl |
map |
-0.02% |
-0.03% |
-0.06% |
-0.06% |
chatGLM2-6b |
lambada openai |
ppl |
17.38 |
17.41 |
17.17 |
|
Llama-2-7b-chat |
Wiki, StackExch, Crawl |
ppl |
3.24 |
3.24 |
3.25 |
|
Stable-Diffusion-V2-1 |
LIAON-5B |
CLIP |
||||
Mistral-7b |
proprietary Mistral.ai |
ppl |
3.29 |
3.47 |
3.49 |
OpenVINO™ Model name |
dataset |
Metric Name |
A, FP32 |
B, FP32 |
C, FP32 |
C, BF16 |
D, FP16 |
|---|---|---|---|---|---|---|---|
bert-base-cased |
SST-2_bert_cased_padded |
0.00% |
0.00% |
0.00% |
-0.09% |
0.00% |
|
bert-large-uncased-whole-word-masking-squad-0001 |
SQUAD_v1_1_bert_msl384_mql64_ds128_lowercase |
F1 |
0.04% |
0.04% |
0.04% |
0.06% |
0.04% |
efficientdet-d0 |
COCO2017_detection_91cl |
coco_precision |
-0.02% |
-0.02% |
-0.02% |
-0.02% |
-0.03% |
mask_rcnn_resnet50_atrous_coco |
COCO2017_detection_91cl_bkgr |
coco_orig_precision |
-0.01% |
-0.01% |
% |
-0.18% |
0.02% |
mobilenet-v2 |
ImageNet2012 |
accuracy @ top1 |
0.00% |
0.00% |
0.00% |
-0.04% |
0.02% |
resnet-50 |
ImageNet2012 |
accuracy @ top1 |
0.02% |
0.02% |
0.00% |
0.01% |
0.01% |
ssd-resnet34-1200 |
COCO2017_detection_80cl_bkgr |
map |
0.00% |
0.00% |
0.00% |
-0.02% |
0.02% |
ssd-mobilenet-v1-coco |
COCO2017_detection_80cl_bkgr |
coco-precision |
0.01% |
0.01% |
0.01% |
0.05% |
-0.03% |
unet-camvid-onnx-0001 |
CamVid_12cl |
mean_iou @ mean |
0.00% |
0.00% |
0.00% |
-0.03% |
-0.03% |
yolo_v3_tiny |
COCO2017_detection_80cl |
map |
% |
0.00% |
0.00% |
0.00% |
-0.02% |
yolo_v8n |
COCO2017_detection_80cl |
map |
0.00% |
0.00% |
0.00% |
0.05% |
-0.03% |
chatGLM2-6b |
lambada openai |
ppl |
17.48 |
17.56 |
17.49 |
||
Llama-2-7b-chat |
Wiki, StackExch, Crawl |
ppl |
3.26 |
3.26 |
|||
Stable-Diffusion-V2-1 |
LIAON-5B |
CLIP |
22.48 |
||||
Mistral-7b |
proprietary Mistral.ai |
ppl |
3.19 |
3.18 |
Notes: For all accuracy metrics except perplexity a “-“, (minus sign), indicates an accuracy drop. For perplexity (ppl) the values do not indicate a deviation from a reference but are the actual measured accuracy for the model.