MAML-Pytorch
MAML-Pytorch copied to clipboard
dragen1860
Elegant PyTorch implementation of paper Model-Agnostic Meta-Learning (MAML)
准确率不变
我使用ALBERT和孪生网络来训练一个主观问题评分模型，训练策略参考的你的代码，孪生网络由双向LSTM和全连接层组成。在训练中，我发现准确率没有提高，一直保持不变。我感觉像是权重没有更新，可能是因为梯度太小导致了权重变化不大。或者，训练策略可能存在问题，但我不确定具体原因。下面是我训练期时的准确率：  ` class MetaTask(nn.Module): def __init__(self, args): super(MetaTask, self).__init__() self.device = 'cuda' if torch.cuda.is_available() else 'cpu' self.loss_fn = nn.CrossEntropyLoss() self.update_lr = args.update_lr self.meta_lr = args.meta_lr self.finetunning_lr = args.finetunning_lr self.n_way...
What is the backup file for and what is the reference navie5 in navie5_train? 
 I trained the model using my own dataset.Does this mean that I have achieved 79.6% accuracy in few-shot classification? But I have only trained for ten minutes. . .I'm...
**__init__ and forward methds for Learner is so complicated** ``` class Learner(nn.Module): """ """ def __init__(self, config, imgc, imgsz): """ :param config: network config file, type:list of (string, list) :param...
第一个问题： 在meta.py文件中  第82行，相当于每次更新一次Meta的参数后，下一次任务的开始，Meta的参数又会被重新初始化，那么我们在for循环结束后更新Meta参数的意义是什么呢？ 第二个问题： 在Meta.py文件中 我们发现有两个“with torch.no_grad():”，那这里面的两个操作的意义是什么呢，感觉并不参与训练，更像是在记录日志
I am writing a [blog](https://metabloggism.github.io/) (I already presented it in this subreddit) and in my [last post](https://metabloggism.github.io/2023/02/07/meta-learning-analysis.html), I did a performance analysis of MAML. I ran several experiments, basically trying...
Hello! When the code is running, training and testing display 6 and 11 numbers respectively, such as training acc: [0.18333333 0.37 0.51333333 0.54666667 0.55333333 0.55 ] Test acc: [0.2974 0.531...
File "C:\xxx\xxx\xxx\xxx\Python39\lib\multiprocessing\reduction.py", line 60, in dump ForkingPickler(file, protocol).dump(obj) AttributeError: Can't pickle local object 'MiniImagenet.__init__..' why?
Metadata
Owner
Metadata
Elegant PyTorch implementation of paper Model-Agnostic Meta-Learning (MAML)