Junwei Deng
Junwei Deng
There has been some input that Triton is used by some users for time series forecasting model serving, we may give such an example to ensure our users that we...
ipex 1.9 does not bring any optimization to chronos's models but it does not mean 1.11(after a large rafactor) it can't benefit chronos models. We need to - [ ]...
We need such a how to guide to help users to understand how to predict future data by using a trained forecaster Critical APIs includes `forecaster.predict` `tsdataset.unscale_numpy()` (and why it...
Chronos: check if learning rate warm up is applied correctly in forecaster's multi-instance training
Nano will apply 'learning rate warmup' automatically with recent updates. While chronos adopt multi-instance training far earlier than this update, we need to check if this affect the accuracy result...
Some nano github action is broken (or very unstable)
1. `time_enc` parameter has weak API guide and makes the whole forecaster workflow inconsistent. 2. autoformerforecaster need `from_tsdataset` function 3. autoformer only returns mse as `val_loss` in evaluation method. 4....
#### Before all This proposal is inspired by the implementation of [GluonTS](https://ts.gluon.ai/api/gluonts/gluonts.dataset.repository.datasets.html). #### Why do we need a Repository Dataset API? - It is inevitable that we have some redundant...
Ray is getting popular for building distributed applications and easy to fit into tsfresh by a `RayDistributor`. # Distributed tsfresh on Ray This repo involves a new `RayDistributor` for tsfresh...
In model.py there are two typos for each run_{} function. ``` enc1.num_samples = 10 enc2.num_samples = 25 ``` should be changed to ``` enc1.num_sample = 10 enc2.num_sample = 25 ```...