Fine-Tuning Tasks

We provide example scripts for few-shot forecasting, imputation and anomaly detection with Timer. The checkpoint is pre-trained and fine-tuned using TSLib.

Supported Tasks

Forecasting: We provide scripts for few-shot forecasting.

Imputation: We adopt segment-level imputation, which is more challenging than point-level imputation.

Anomaly Detection: We build a benchmark using UCR Anomaly Archive. The task aims to predict normal future series and detect anomalies in advance.

We provide the README files illustrating each task under the folder ./scripts/.

Code for Fine-tuning

1. Use Python 3.10 and install necessary dependencies.

pip install -r requirements.txt

2. Put downstream datasets from Google Drive or Baidu Drive under the folder ./dataset/.

3. Put the checkpoint from Google Drive or Baidu Drive under the folder ./checkpoints/.

4. Train and evaluate the model. We provide the above tasks under the folder ./scripts/.

# forecasting
bash ./scripts/forecast/ECL.sh

# segement-level imputation
bash ./scripts/imputation/ECL.sh

# anomaly detection
bash ./scripts/anomaly_detection/UCR.sh

5. We provide detailed task descriptions on corresponding README files: [Forecasting] [Imputation] [Anomaly Detection]