Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Matrix multiplication error during training TFT

See original GitHub issue

Environment Details

Python version: 3.7.10
Operating System: Google Colab web platform

Error Description

I get a matrix multiplication error during training TFT

RuntimeError: mat1 and mat2 shapes cannot be multiplied (4608x2 and 1x32)

Steps to reproduce

Open Colab https://colab.research.google.com/
Run the code

!pip install pytorchts -q
!curl https://forecasters.org/data/m3comp/M3C.xls --create-dirs -o /root/.mxnet/gluon-ts/datasets/M3C.xls 

from gluonts.dataset.repository.datasets import get_dataset
from pts.model.tft import TemporalFusionTransformerEstimator
from pts import Trainer

dataset = get_dataset("m3_monthly", regenerate=False)

estimator = TemporalFusionTransformerEstimator(
    freq=dataset.metadata.freq,
    prediction_length=dataset.metadata.prediction_length,
    context_length=dataset.metadata.prediction_length,
    dropout_rate=0.1,
    num_outputs=15,
    trainer=Trainer(device='cpu',
                    epochs=20,
                    learning_rate=1e-3,
                    num_batches_per_epoch=100,
                    batch_size=128))

predictor = estimator.train(dataset.train)

Issue Analytics

State:
Created 2 years ago
Comments:13

Top GitHub Comments

1reaction

kashifcommented, Jun 2, 2021

@nsoma97 @vpozdnyakov sorry i didnt get time to look into it… I will try to squeeze it in this week… the issue most probably is that I have some value hard-coded from the dataset I was testing with…

0reactions

vpozdnyakovcommented, Jul 7, 2021

it works, thank you