@karpathy
high engagement
X post
#Neural architectures for time series
Explained how transformer-based models outperform LSTMs for long-horizon time series forecasting by capturing global dependencies in sequential data.
Related Jobs
Opportunities for Predictive Modelling
More Sources for Predictive Modelling
@jeremyphoward
Shared a fast.ai case study on using neural nets for retail demand forecasting that reduced inven...
#Real-world forecasting applications
3 media
@TensorFlow
Posted a thread with tutorials and notebooks on building probabilistic forecasting models with Te...
#Learning resources and tutorials