A-cloud-bit

About the testing

4

Hello author, the code and ideas in this paper are very well written. I have a question, using azimuth angles of 0, 1, and 2 as the standard for saving...

about Command '['ninja', '-v']' returned non-zero exit status 1.

1

Error building extension '_pvcnn_backend': [1/13] /usr/bin/nvcc I have modified Gcc according to the requirements, but this error still persists. It's very painful and I don't know how to solve it

博主你好，我想问一下，在这里从验证集取数据，对模型进行微调，但是我们并没有更新元模型权重。这里的目的是什么呢？只是说单纯的验证一下模型的性能吗？ for support_images, support_labels, query_images, query_labels in val_loader: # Get variables support_images = support_images.float().to(dev) support_labels = support_labels.long().to(dev) query_images = query_images.float().to(dev) query_labels = query_labels.long().to(dev) loss, acc = maml_train(model, support_images, support_labels, query_images,...

关于标签分配问题

1

img_dirs = random.sample(self.file_list, self.n_way) for label, img_dir in enumerate(img_dirs): print('此次的标签为：',label) img_list = [f for f in glob.glob(img_dir + "**/*.png", recursive=True)] images = random.sample(img_list, self.k_shot + self.q_query) 据我所知，这个火星数据集大约有1000多个类别，但是这里分配真实标签的时候似乎没有什么依据，而是按照索引顺序分配的。关键是，这个标签数值分配似乎是错误的。只能分配0-5

reptile不需要划分支持集和查询及，但是你这里读取了支持集和查询及，此外为什么对于每个任务训练的时候 ，又从支持集读取了10张图片呢？这到底是5way 几shot呢？ override = self.inner_optimizer if is_train else None #其实是相当于创建了一个副本，这个优化器和网络的更新都不会影响原来的 with higher.innerloop_ctx(self.network, self.inner_optimizer, track_higher_grads=False) as (fmodel, diffopt): for step in range(self.args.n_inner): if is_train: index = np.random.permutation(np.arange(len(test_input)))[:10] train_input = test_input[index]...

Very good job, we need the number of parameters for the model now. Could you please let us know

ser.add_argument('--classes', default=5, type=int, help='classes in base-task (N-way)') parser.add_argument('--shots', default=5, type=int, help='shots per class (K-shot)') parser.add_argument('--train-shots', default=10, type=int, help='train shots') parser.add_argument('--batch', default=10, type=int, help='minibatch size in base task') 我不理解的是每个任务是n-ways--k-shot但是你有两个参数。此外，这个minibatch size in...

Delving into Inter-Image Invariance for Unsupervised Visual Representations Code

### The doc issue Delving into Inter-Image Invariance for Unsupervised Visual Representations Code ### Suggest a potential alternative/fix _No response_