CVSE:用于图像-文本匹配的共识感知视觉语义嵌入论文的官方源代码（ECCV 2020）
需积分: 13 44 浏览量
2021-04-12
20:32:22
上传
评论 2
收藏 214.1MB ZIP 举报
CVSE-master.zip （64个子文件） 
CVSE-master runs f30k CVSE_f30k log 
events.out.tfevents.1608974921.DESKTOP-VNJSTH5 40B model_best.pth.tar 96.57MB coco CVSE_COCO model_best.pth.tar 98.32MB figures 
framework_CVSE.jpg 472KB train_coco.py 14KB data f30k_annotations Concept_annotations f30k_concepts_glove_word2vec.pkl 703KB category_concepts.json 4KB all_f30k_concept_label.json 2.95MB f30k_adj_concepts.pkl 1.03MB vocab_f30k_concepts.pkl 6KB f30k dataset_flickr30k.json 36.54MB coco_annotations Concept_annotations category_concepts.json 4KB coco_adj_concepts.pkl 1.03MB trainval_concept_label.json 16.01MB vocab_trainval_concepts.pkl 6KB coco_concepts_glove_word2vec.pkl 703KB coco_to_f30k_annotations Concept_annotations category_concepts.json 4KB Flickr30k_test_concept_label.json 89KB coco_adj_concepts.pkl 1.03MB vocab_trainval_concepts.pkl 6KB coco_concepts_glove_word2vec.pkl 703KB coco annotations coco_test_ids.npy 195KB coco_restval_ids.npy 1.16MB coco_dev_ids.npy 195KB coco_train_ids.npy 3.16MB captions_train2014.json 63.69MB image_info_test2014.json 7.5MB captions_val2014.json 30.92MB download.sh 402B model_CVSE.py 36KB __pycache__ model_CVSE.cpython-36.pyc 21KB evaluation.cpython-36.pyc 13KB data.cpython-36.pyc 11KB vocab.cpython-36.pyc 3KB evaluation.py 19KB evaluate.py 4KB .idea $PRODUCT_WORKSPACE_FILE$ 461B $CACHE_FILE$ 467B misc.xml 197B vcs.xml 180B modules.xml 308B dictionaries 163B workspace.xml 7KB Code_github_ECCV2020_revised.iml 453B inspectionProfiles profiles_settings.xml 128B vocab coco_vocab.pkl 316KB f30k_vocab.json 293KB f8k_precomp_vocab.pkl 83KB f8k_vocab.pkl 88KB coco_precomp_vocab.pkl 268KB coco_precomp_vocab.json 392KB f30k_vocab.pkl 230KB f30k_precomp_vocab.json 14B f30k_precomp_vocab.pkl 227KB train_f30k.py 14KB README.md 5KB util utils.py 4KB __pycache__ C_GCN.cpython-36.pyc 4KB util_C_GCN.cpython-36.pyc 3KB utils.cpython-36.pyc 4KB util_C_GCN.py 6KB C_GCN.py 4KB data.py 18KB vocab.py 3KB资源评论
