Time series Timeseries Deep Learning Machine Learning Python Pytorch fastai | State-of-the-art Deep Learning library for Time Series and Sequences in Pytorch / fastai
Originally posted by **Samcro5C** November 30, 2022
When I tried to use the fastai callback [TensorBoardCallback](https://docs.fast.ai/callback.tensorboard.html#tensorboardprojectorcallback) for a TSForecaster (minimal example from Readme), I ran into a RuntimeError (_RuntimeError: Exception occured in `TensorBoardCallback` when calling event `after_epoch`: Can't call numpy() on Tensor that requires grad. Use tensor.detach().numpy() instead._), so I was wondering if tsai supports tensorboard at all or if this is a bug. Thank you in advance.
Discussed in https://github.com/timeseriesAI/tsai/discussions/633
Code to reproduce the issue:
Here's the full traceback: