Alibaba-MIIL/ImageNet21K

Model weights trained on Stanford cars

ma-siddiqui opened this issue · 5 comments

ma-siddiqui commented

Hi,

Can you please share the link to the weights file of model trained on Stanford cars data set. I am unable to get the expected results using https://miil-public-eu.oss-eu-central-1.aliyuncs.com/model-zoo/ImageNet_21K_P/models/tresnet_l_v2_miil_21k.pth

Please guide.

Thanks,

ma-siddiqui commented

fortunately found here. Please correct me if this is not a correct link

https://github.com/Alibaba-MIIL/TResNet

ma-siddiqui commented

Thank you. What is the correct model architecture for these weights? Should I use imageNet21K or original TresNet repo. Actually I faced some errors using the code TresNet repo.

mrT23 commented

if you faced errors, give full reproduction code. use the model LTResNet-V2 form tresnet repo

ma-siddiqui commented

Hi,

I am facing below errors

File "TResNet/infer.py", line 44, in
main()
File "TResNet/infer.py", line 26, in main
model = create_model(args).cuda()
File "TResNet/src/models/utils/factory.py", line 23, in create_model
model = TResnetL_V2(model_params)
File "TResNet/src/models/tresnet_v2/tresnet_v2.py", line 239, in TResnetL_V2
model = TResNetV2(layers=layers_list, num_classes=num_classes, in_chans=in_chans,
File "TResNet/src/models/tresnet_v2/tresnet_v2.py", line 160, in init
layer2 = self._make_layer(Bottleneck, self.planes * 2, layers[1], stride=2, use_se=True,
File "TResNet/src/models/tresnet_v2/tresnet_v2.py", line 215, in _make_layer
layers.append(block(self.inplanes, planes, stride, downsample, use_se=use_se,
File "TResNet/src/models/tresnet_v2/tresnet_v2.py", line 117, in init
anti_alias_layer(channels=planes, filt_size=3, stride=2))
TypeError: init() got an unexpected keyword argument 'remove_model_jit'

Thanks,