All Questions
88 questions
0
votes
1
answer
68
views
Unable to return a boolean variable from Pytorch Dataset's __get_item__
I have a pytorch Dataset subclass and I create a pytorch DataLoader out of it. It works when I return two tensors from DataSet's __getitem__() method. I tried to create minimal (but not working, more ...
- 3,007
1
vote
0
answers
427
views
How to use parallel torch cuda streams without causing oom? (example included)
I'm storing a large amount of tensors data in a cpu memory, and the intended workflow is to process them using the GPU. And while one chunk is being processed, simultaneously transfer the previous ...
- 4,630
1
vote
1
answer
156
views
Pytorch `DataSet.__getitem__()` called with `index` bigger than `__len__()`
I have following torch dataset (I have replaced actual code to read data from files with random number generation to make it minimal reproducible):
from torch.utils.data import Dataset
import torch
...
- 25k
1
vote
0
answers
29
views
PyTorch's Linear layer not training
I am trying to implement a Deep Embedded Self-Organizing Map (DESOM) which is an Autoencoder together with a trainable SOM as a trainable layer which I implement using a Linear layer:
class SOM(nn....
- 2,498
0
votes
2
answers
1k
views
Huggingface pipeline available models
I'm working with Huggingface in Python to make inference with specific LLM text generation models. So far I used pipelines like this to initialize the model, and then insert input from a user and ...
- 5,243
1
vote
2
answers
496
views
Gradients for selection from array operation in PyTorch
This is a followup question from here. I have obtained a tensor, say, d with gradients. Now I have another tensor array, say e from which I need to pick the first d elements. MWE below.
import torch
...
- 3,482
2
votes
2
answers
413
views
Pass gradients through min function in PyTorch
I have a variable, say, a which has some gradient associated with it from some operations before. Then I have integers b and c which have no gradients. I want to compute the minimum of a, b, and c. ...
- 3,482
0
votes
1
answer
37
views
LSTM layer followed by BatchNorm Layer error in dimensions
I am trying to train a model with the following architecture :
self.lstm1 = nn.LSTM(in_channels, hidden_channels, num_layers, batch_first=True, dropout=dropout_prob)
self.batchnorm1 = nn.BatchNorm1d(...
1
vote
1
answer
919
views
Calculating the Fisher metric in PyTorch
Given some simple PyTorch model, how do I compute the Fisher Metric?
Here's a (useless for practical purposes) trivial model which uses a single linear layer to solve the matrix equation Ax=b, where A ...
1
vote
1
answer
1k
views
how to rename classes of trained model in Pytorch
I've trained my object detection model based on YOLOV7 and YOLOV5. Now, for some reason i need to change name of classes.
Is it possible to change classes names and save it again as Pytorch ML model ...
- 969
0
votes
0
answers
30
views
while installing dependencies from the repo its giving me error
here are the dependencies required to run this simple repository on my mac book for tiny face detection in python and it's giving me errors while installing the requirements.txt file
here is the error ...
0
votes
1
answer
344
views
Got TypeError when adding return_indices=True to nn.MaxPool2d in pytorch
I am New to pyotch, i am trying to create an autoencoder in pytorch, here is my code
The encoder:
# B = Batch size
# encoder (B, 3, 224, 224) => (B, 8)
class Encoder(nn.Module):
def __init__(...
- 1
2
votes
1
answer
371
views
Saving the model architecture with activation functions in PyTorch
I use PyTorch for training neural networks. While saving the model, the weights of the network are saved, while the activation functions are not captured. Now, I reload the model from the saved ...
- 3,482
3
votes
3
answers
4k
views
KeyError: 'eval_loss' in Hugginface Trainer
I am trying to build a Question Answering Pipeline with the Hugginface framework but facing the KeyError: 'eval_loss' error. My goal is to train and save the best model at last and evaluate the ...
- 12.7k
0
votes
1
answer
233
views
RuntimeError: mat1 and mat2 shapes cannot be multiplied (1x7 and 1x1)
I faced this issue when I tried to train the model. This model is meant to predict Celsius to Fahrenheit.
Do help me to review my code below:
# this file is a simple example of how to use pytorch to ...
- 365