All Questions
Tagged with pytorch pytorch-lightning
471 questions
0
votes
1
answer
7
views
RecursionError when using Opacus PrivacyEngine with PyTorch Lightning: maximum recursion depth exceeded
I'm implementing a differentially private recommendation system using PyTorch Lightning and Opacus, but I'm encountering a RecursionError during training. Here's my setup:
Problem
When I run my ...
- 45
1
vote
1
answer
39
views
Pytorch Lightning logs separately for train, validation and test datasets
I am trying to log the loss and auc for all 3 of my datasets - train, validation and test.
The datamodule defines the 3 loaders and I finally invoke the model as:
trainer.fit(model,datamodule)
trainer....
- 173
0
votes
0
answers
42
views
How to wrap a torch.jit model inside a torch Module?
I'm trying to call a TorchScript model inside a torch.nn.Module but got an error related to pickle.
Here's the code to reproduce:
import torch
import torch.nn as nn
# A simple base model to create a ...
- 357
0
votes
1
answer
17
views
How to apply min-max scaling on a IterableDataset?
I'm using an iterableDataset because I have massive amounts of data. And since IterableDataset does not store all data in memory, we cannot directly compute min/max on the entire dataset before ...
- 15
2
votes
3
answers
130
views
How does Hydra `_partial_` interact with seeding
In the configuration management library Hydra, it is possible to only partially instantiate classes defined in configuration using the _partial_ keyword. The library explains that this results in a ...
- 1,232
0
votes
0
answers
42
views
Lightning Tensorboard Hparams tab not showing custom metric
I'm training a neural network built with pyTorch Lightning and I'm trying to have the HParams tab working in tensorboard.
Following the official outdated guide I initialize the logger with ...
- 1,892
0
votes
0
answers
44
views
how to forecast future after training pytorch time series model?
I am learning multivariate time series forecasting using pytorch lightning. The book always check model prediction with testing dataset. The book never taught me how to make prediction for future, ...
- 4,280
1
vote
0
answers
81
views
mlflow.pytorch.load_model failing due to failure "No module named 'src.<mymodelname>'" issue with unpickler.load
I am working to support some model developers by prototyping some functionality with MLflow model registry.
We successfully register versions of the model fine (it's a Pytorch Lightning model), and I ...
- 3,896
0
votes
0
answers
36
views
DistNetworkError when using multiprocessing_context parameter in pytorch dataloader
Because of some special reasons I want to use spawn method to create worker in DataLoader of Pytorch, this is demo:
import torch
import torch.nn as nn
import torch.optim as optim
from torch.utils.data ...
- 1,025
0
votes
0
answers
122
views
Lightning + MLFlowLogger logging checkpoints, can't register in Model Registry
I am using a Sagemaker-managed MLflow tracking server, and Pytorch LightningCLI --
The tracking server was created with "AutomaticModelRegistration": true, and we're using the MLFlowLogger ...
- 3,896
0
votes
0
answers
21
views
Using weight in TimeSeriesDataSet
I have a problem how to use weights inside the TimeSeriesDataSet. I have error ValueError: tensor.ndim=2 > like.ndim=0.
Any help?
I have a dataset, where for each item I have 66 time steps. There ...
- 1
1
vote
0
answers
56
views
Storing the dataset on CPUs but training on GPUs with Lightning
I'm having a bit of an issue when the dataloader loads the training and validation sets into the GPU and it runs out of RAM causing the program to terminate. I was recommended to set the device to CPU ...
- 59
0
votes
0
answers
37
views
Faster RCNN with Pytorch Lightning not showing better results mAP
I've tried to fine-tune the model for binary class object detection
Following is the code
There are only single class in COCO dataset with label 1 to classify that as object.
Tensorboard showing poor ...
- 10.4k
0
votes
1
answer
686
views
view size is not compatible with input tensor's size and stride
I'm trying to training F-RCNN based on coco dataset on my images. Image size is 512X512
I've tested dataloader separately and it works and prints the batch images and BB details
Also i've tried to ...
- 10.4k
0
votes
0
answers
19
views
PyTorch Bounding Box in COCO Format resize issue: list indices must be integers or slices, not tuple [duplicate]
I'm trying this example
https://lightning.ai/lightning-ai/studios/object-detection-with-pytorch-lightning#full-training-example
And this is the output code of bounding box in the annotation of image ...
- 10.4k