facebookresearch/SlowFast

Runtime error

Open

#172 opened on Apr 20, 2020

View on GitHub
 (0 comments) (0 reactions) (0 assignees)Python (6,064 stars) (1,167 forks)batch import
help wanted

Description

Thank you very much for your excellent work. I wanted to train with the ava data set, but it went wrong: RuntimeError: Caught RuntimeError in DataLoader worker process 0. RuntimeError: invalid argument 0: Sizes of tensors must match the except in dimension 0. Got 3 and 2 dimension in the at 1 / pytorch aten/SRC/TH/generic/THTensor CPP: 689 My command line is: python tools/run_net.py --cfg configs/AVA/c2/SLOWFAST_32x2_R101_50_50.yaml DATA.PATH_TO_DATA_DIR /data/ava/videos NUM_GPUS 4 TRAIN.BATCH_SIZE 8 Is there anything wrong with my data set path setting?How should you modify it?Thank you very much! My data set and configuration files were downloaded according to your command.

Contributor guide

Tech stack
pythonpytorch
Domain
machine learning
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
needs investigation
Prerequisites
PythonPyTorchSlowFast setupAVA dataset
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
Investigate the data loading pipeline in SlowFast, specifically the AVA dataset configuration. Check the config file referenced (configs/AVA/c2/SLOWFAST 32x2 R101 50 50.yaml) and verify the data directory structure matches expectations. Look into the DataLoader worker error to identify tensor size mismatch, likely due to incorrect data preprocessing or file paths. Consult the repository documentation for dataset setup instructions. Since the issue has no comments or PRs, no prior investigation is visible.