Stack Overflow
Stack Overflow

Questions tagged [meta-learning]

18 questions
7
votes
0 answers

How does one create a distributed data loader with PyTorch's TorchMeta for meta-learning?

I was trying to create a pytorch distributed data laoder with torchmeta but it failed with a deadlock: python ~/ultimate-utils/tutorials_for_myself/my_torchmeta/torchmeta_ddp.py test_basic_ddp_example ABOUT TO SPAWN WORKERS (via mp.spawn) ->…
Charlie Parker
  • 5,884
  • 57
  • 198
  • 323
1
vote
1 answer

Meta-learning to find optimal model from pre-trained models in Tensorflow

I have many pre-trained models with a different number of layers (Models are not Sequential). Training data had a shape (1, 1, 103) for these models and output was a class label between 0 and 9. I loaded these saved models, set all layers as…
Karan Owalekar
  • 947
  • 10
  • 31
1
vote
1 answer

When should one call .eval() and .train() when doing MAML with the PyTorch higher library?

I was going through the omniglot maml example and saw that they have net.train() at the top of their testing code. This seems like a mistake since that means the stats from each task at meta-testing is shared: def test(db, net, device, epoch, log): …
Charlie Parker
  • 5,884
  • 57
  • 198
  • 323
0
votes
0 answers

How to use the trained model in meta learning

I have trained a meta learning model using learn2learn in python. I train the model using 5 way 5 shot. After training, can I use the model to classify a 395 class dataset? I train the model using 5 way 5 shot. But I think the result can only use to…
林佑宸
  • 1
0
votes
1 answer

Is this the correct implementation of a MAML model?

I have used CLIP embeddings of image and text as the input and the output is a label ranging from 0 to 5 (6 way label). I tried to make an implemention of this multimodal 6 way classification using meta learning. I tried using a code which includes…
varun80042
  • 1
0
votes
0 answers

Improving Performance of One Shot Learning Architecture for 1D Arrays in TensorFlow

The objective is to compare two 1D array signals. The approach involves converting the input signals into their corresponding encodings, followed by a distance calculation. The anticipated output is a prediction value in proximity to 0 if the…
Karan Owalekar
  • 947
  • 10
  • 31
0
votes
0 answers

How to update models based on the updated params saved in sollections.OrderedDict()?

I am coding the MAML meta-learning model. I am not sure if I am doing correctly or not? After training on support set and calculating the loss over query set, I am updating the model parameters by sending the loss to a function defined in the meta…
S.EB
  • 1,966
  • 4
  • 29
  • 54
0
votes
0 answers

Image Segmentation using MAML algorithm (same objects exist in all tasks)

I have an n-takes k-shots medical image segmentation problem. -Tasks: Different human organs ex: liver, spleen, kindness etc... -Shots: 10 CT scans NIFTI images, where all tasks(human organs) exist in all shots, but one of them is labelled to match…
AMAS AL
  • 13
  • 3
0
votes
0 answers

Applying meta-learning for neural machine translation

Does anyone have any idea of how to apply meta-learning for neural machine translation? I have read some papers that consider multiple language pairs as different tasks for meta-learning. Is it possible to do it just with one pair of language? for…
Kyrsten
  • 1
  • 1
0
votes
1 answer

Too much fluctuation in F1 Score curve during meta training with MAML

I am training VGG11 on a custom image dataset for 3-way 5-shot image classification using MAML from learn2learn. I am encapsulating the whole VGG11 model with MAML, i.e., not just the classification head. My hyperparameters are as follows: Meta LR:…
The Exile
  • 644
  • 8
  • 14
0
votes
0 answers

ValueError: Graph disconnected: cannot obtain value for tensor KerasTensor(type_spec=TensorSpec(shape=(None, None, None, 3)

I am working on a MetaLearning approach to solving an image segmentation problem. I trained 7 different models on different classes related to the eye dataset provided under Oculus SBVPI (iris, pupil, sclera, canthus, periocular, vessels,…
user8516309
0
votes
0 answers

How to combine multiple dataset efficiently to solve using meta learning?

I am solving a meta-learning problem using Reptile Algorithm as used here. I have two datasets. One contains the following classes: iris, pupil, and sclera along with their annotations. Another contains classes as follows: iris, pupil, sclera, and…
Na462
  • 11
  • 2
0
votes
0 answers

Vanishing parameters in MAML JAX (Meta Learning)

I am working on an implementation of MAML (see https://arxiv.org/pdf/1703.03400.pdf) in Jax. When training on a distribution of simple linear regression tasks it seems to perform fine (takes a while to converge but ultimately works). However when…
Sefton de Pledge
  • 19
  • 4
0
votes
1 answer

How to split classes in few-shot classification using CIFAR-10?

I want to train a model that perform a few-shot image classification using CIFAR-10. So I have to train the model with a small amount of classes and use the rest of the classes for the testing. I'm wondering if I have only 10 classes, how can i do…
Giulia
  • 11
  • 4
0
votes
1 answer

Why is RandomCrop with size 84 and padding 8 returning an image size of 84 and not 100 in pytorch?

I was using the mini-imagenet data set and noticed this line of code: elif data_augmentation == 'lee2019: normalize = Normalize( mean=[120.39586422 / 255.0, 115.59361427 / 255.0, 104.54012653 / 255.0], …
Charlie Parker
  • 5,884
  • 57
  • 198
  • 323
1
2