peteanderson80 / bottom-up-attention

Bottom-up attention model for image captioning and VQA, based on Faster R-CNN and Visual Genome
http://panderson.me/up-down-attention/
MIT License
1.43k stars 378 forks source link

Could not open test.prototxt #71

Open chlorane opened 5 years ago

chlorane commented 5 years ago

I'm using the res-101 network and tools/generate_tsv.py to extract bounding box features to a tab-separated-values (tsv) file. I'm using a new dataset based on cartoon series. So I run: python tools/generate_tsv.py --cfg experiments/cfgs/faster_rcnn_end2end_resnet.yml --def models/vg/ResNet-101/faster_rcnn_end2end/test.prototxt --out test2014_resnet101_faster_rcnn_genome.tsv --net data/faster_rcnn_models/resnet101_faster_rcnn_final.caffemodel --split cartoon

However, when I tried to run generate_tsv.py, an error happened: Process Process-1: Traceback (most recent call last): File "/home/chlorane/anaconda3/lib/python3.7/multiprocessing/process.py", line 297, in _bootstrap self.run() File "/home/chlorane/anaconda3/lib/python3.7/multiprocessing/process.py", line 99, in run self._target(*self._args, **self._kwargs) File "tools/generate_tsv.py", line 162, in generate_tsv net = caffe.Net(prototxt, caffe.TEST, weights=weights) RuntimeError: Could not open file ./models/vg/ResNet-101/faster_rcnn_end2end/test.prototxt

Even if I use the absolute path here, the error still happens. What's wrong and how can I solve it?

ghost commented 5 years ago

Hi, I had the same problem. You can solve it by change the arges --def models/vg/ResNet-101/faster_rcnn_end2end/test.prototxt to --def models/vg/ResNet-101/faster_rcnn_end2end_final/test.prototxt. That's work for me!

alice-cool commented 3 years ago

https://github.com/peteanderson80/bottom-up-attention/issues/71#issuecomment-542163318

Dear scholar, did you run the result? I got all zero and NAN in the second picture. I am so cunfused

image image