Google has released TimesFM 2.0, a pre-trained model for time series forecasting, a major update designed to improve forecast accuracy and promote the field of artificial intelligence. TimesFM 2.0 is shared through open source and is convenient for researchers and developers to use. The model can handle univariate time series up to 2048 time points, supports any forecast time span, and has strong forecasting capabilities. Its training data covers multiple fields, such as energy, transportation, etc., providing a rich learning basis for the model. The emergence of TimesFM 2.0 will significantly improve the efficiency and accuracy of time series forecasting and provide more powerful data analysis tools for all walks of life.
The TimesFM2.0 model has powerful functions and can handle univariate time series forecasts up to 2048 time points, and supports any forecast time span.
It is worth noting that although the model is trained with a maximum context length of 2048, in practical applications, longer contexts can be processed. The model focuses on point prediction, while 10 quantile heads are experimentally provided, but these have not been calibrated after pre-training.
In terms of data pre-training, TimesFM2.0 contains a combination of multiple data sets, including the pre-training set of TimesFM1.0 and additional data sets from LOTSA. These data sets cover multiple fields, such as residential electricity load, solar power generation, traffic flow, etc., providing a rich foundation for model training.
Through TimesFM2.0, users can more easily conduct time series predictions and promote the development of various applications, including retail sales, stock trends, website traffic and other scenarios, environmental monitoring and intelligent transportation and other fields.
Model entrance: https://huggingface.co/google/timesfm-2.0-500m-pytorch
Highlight:
TimesFM2.0 is a new time series prediction model launched by Google, focusing on improving the accuracy of time series prediction.
The model supports forecasts up to 2048 time points and can handle any forecast time span.
Users can freely choose the prediction frequency based on different time series characteristics to improve the flexibility of prediction.
All in all, TimesFM 2.0, with its powerful functions and ease of use, will bring new breakthroughs in the field of time series forecasting and be widely used in various industries. Looking forward to its development and application in the future.