Resnet50_test / metadata.yaml
RoyalJ's picture
Upload quantized model and evaluation results
d01fde5
raw
history blame contribute delete
239 Bytes
model_name: Resnet50
liteml_config: 'liteml_config.yaml'
quantization_type: W8A8 static
dataset: ImageNet
notes: Quantized W8A8 Resnet50 model from torchvision
# TODO: move float accuracy somewhere else
top1_float: 0.555
top5_float: 0.777