def i2t()

in senteval/tools/ranking.py [0:0]

• 31 lines of code
• 6 McCabe index (conditional complexity)

    def i2t(self, images, captions):
        """
        Images: (5N, imgdim) matrix of images
        Captions: (5N, sentdim) matrix of captions
        """
        with torch.no_grad():
            # Project images and captions
            img_embed, sent_embed = [], []
            for i in range(0, len(images), self.batch_size):
                img_embed.append(self.model.proj_image(
                    Variable(images[i:i + self.batch_size])))
                sent_embed.append(self.model.proj_sentence(
                    Variable(captions[i:i + self.batch_size])))
            img_embed = torch.cat(img_embed, 0).data
            sent_embed = torch.cat(sent_embed, 0).data

            npts = int(img_embed.size(0) / 5)
            index_list = []

            ranks = np.zeros(npts)
            for index in range(npts):

                # Get query image
                query_img = img_embed[5 * index]

                # Compute scores
                scores = torch.mm(query_img.view(1, -1),
                                  sent_embed.transpose(0, 1)).view(-1)
                scores = scores.cpu().numpy()
                inds = np.argsort(scores)[::-1]
                index_list.append(inds[0])

                # Score
                rank = 1e20
                for i in range(5*index, 5*index + 5, 1):
                    tmp = np.where(inds == i)[0][0]
                    if tmp < rank:
                        rank = tmp
                ranks[index] = rank

            # Compute metrics
            r1 = 100.0 * len(np.where(ranks < 1)[0]) / len(ranks)
            r5 = 100.0 * len(np.where(ranks < 5)[0]) / len(ranks)
            r10 = 100.0 * len(np.where(ranks < 10)[0]) / len(ranks)
            medr = np.floor(np.median(ranks)) + 1
            return (r1, r5, r10, medr)