self-critical.pytorch: evaluation error using pre-trained model

Hi Ruotian,

I tried to run the following script: python eval.py --model resnet50.pth --infos_path infos.pkl --image_folder ./image/val2014_coco/ --num_images 1

The model was resnet50 (and resnet101) downloaded from your google drive. But I got the error: Traceback (most recent call last): File "eval.py", line 102, in <module> model.load_state_dict(torch.load(opt.model)) File "/home/wentong/miniconda2/lib/python2.7/site-packages/torch/nn/modules/module.py", line 522, in load_state_dict .format(name)) KeyError: 'unexpected key "conv1.weight" in state_dict'

I have searched online but there was little information about that. I guess you have used multiple gpus.

Any advice? Thank you for your implementation.