resnet50-int8-ov
OpenVINO IR export of torchvision/resnet50 (IMAGENET1K_V2), quantized to INT8 — 25 MB.
| Quantization | weight compression (NNCF) |
| Lesson | 01 Hello image classification |
| Built by | convert/convert_all.py of the ARCademy OpenVINO courseware |
Static input [1,3,224,224]. Preprocessing (resize 224, ImageNet normalisation) happens in the lesson code.
from huggingface_hub import snapshot_download
model_dir = snapshot_download("circulus/resnet50-int8-ov")
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support