Transfer learning for Time Series Forecasting

Transfer learning refers to the process of pre-training a flexible model on a large dataset and using it later on other data with little to no training. It is one of the most outstanding 🚀 achievements in Machine Learning 🧠 and has many practical applications.

For time series forecasting, the technique allows you to get lightning-fast predictions ⚡ bypassing the tradeoff between accuracy and speed.

This notebook shows how to generate a pre-trained model and store it in a checkpoint to make it available for public use to forecast new time series never seen by the model. If you want to see a proof of concept in action you can visit this Demo.

If you want to use our Low Latency API for forecasting you can SingUp here.

You can contribute with your pre-trained models by following this Notebook and sending us an email at federico[at]nixtla.io

You can also take a look at list of pretrained models here. Currently we have this ones avaiable in our API or Demo. You can also download the .ckpt:

• Pretrained N-HiTS M4 Hourly
• Pretrained N-HiTS M4 Hourly (Tiny)
• Pretrained N-HiTS M4 Daily
• Pretrained N-HiTS M4 Monthly
• Pretrained N-HiTS M4 Yearly
• Pretrained N-BEATS M4 Hourly
• Pretrained N-BEATS M4 Daily
• Pretrained N-BEATS M4 Weekly
• Pretrained N-BEATS M4 Monthly
• Pretrained N-BEATS M4 Yearly

See how to load and use the pretrained models in this Notebook. If you want us to include more model or train some private ones in your own data, contact us at: federico@nixtla.io.

If you are interested in the transfer learning literature, take a look at this paper:

• Meta-learning framework with applications to zero-shot time-series forecasting
• N-HiTS: Neural Hierarchical Interpolation for Time Series Forecasting