Haojun Jiang(蒋昊峻)
@pqviet Hi, I send it to you. Please check your email.
Yes. Obviously, there is a bug in function "extract_patch". def extract_patch(dataset): extracted = [] for img in **train_dat**: img_ = img[0].reshape(96,96,1) extracted.append(img_) img_ = img[1].reshape(96,96,1) extracted.append(img_) return np.array(extracted)
Just as you cited, "Flickr30K Entities [38] augments the original Flickr30K [58] with short region phrase correspondence annotations." which means the original sentences of Flickr30K are splited to short phrases...
Did you set smaller learning rate? Since you train the model on one GPU, the batchsize is only 1/8 of the original. Thus, the original learning is too large for...
On Something-something dataset, we found that fG and fL do not have to correspond strictly. In Table 3, TSM+ uses 8 frames for both fL network. In contrast, we reduce...
@ZhanYang-nwpu Hi, I send it to you. Please check your email. Please leave a STAR ⭐ if this project helps your research!
Do you have any idea about where the gap comes from?
@evelinehong Hi, thanks for sharing an interesting paper. Please pull a request. Here is the tutorial of [contributing to others projects](https://docs.github.com/en/get-started/quickstart/contributing-to-projects). First, think about which category the work should belong...
@OhadRubin Hi, did you try to use the principle initialization method？And did this work？
Did you solve the problem？