Practical overview
When it fits
Use it when you want to work directly with Python scripts and experiment with stock time-series models, rather than consume forecasts from a hosted service.
Who it is for
- Developers adapting a script-based stock-forecasting workflow.
- Researchers comparing neural-network approaches on stock time-series data.
Documented uses
- Retrieve and preprocess market data for a forecasting experiment.
- Train a selected neural-network model.
- Generate a multi-day forecast using trained weights and matching configuration.
- Evaluate forecasting results on independent samples.
Requirements & limits
The repository is public and licensed under GPL-3.0.
- Python 3.10 and an isolated environment are recommended; project dependencies must be installed.
- Data acquisition relies on an external provider, and some architectures have additional data, configuration or resource requirements.
- Prediction requires matching model weights and configuration, including normalization parameters.
- The project describes forecasts as research outputs and calls for independent validation, data-rights checks and evaluation under real trading costs.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Train a Transformer and request a forecast
Follow the documented example workflow from market-data retrieval through a three-day prediction.
Before you start
- Python 3.10 is recommended, using a dedicated conda environment or equivalent virtual environment.
- Install the project dependencies from requirements.txt.
- The data-retrieval example uses an external data source; its interface may change.
Steps
Retrieve example market data with the command below.
python scripts/getdata.py --api akshare --code 000001.SZPreprocess the data with the command below.
python scripts/data_preprocess.py --pklname train.pklTrain a Transformer with the command below.
python scripts/train.py --mode train --model transformer --epoch 2Request a three-day prediction with the command below, using the corresponding saved model weights and configuration.
python scripts/predict.py --model transformer --test_code 000001 --predict_days 3
Expected outcome
A prepared dataset, trained Transformer model and requested three-day forecast for the specified test code.
Guide limitations
- The data interface can fail if the upstream service changes.
- Check data authorization, adjustment method and training/test time boundaries.
- Forecasts are for research; evaluate on independent samples and account for real trading costs.
Compare options for Model inference and serving
These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.
| Project | Documented purpose | Platform | Access and requirements |
|---|---|---|---|
| Stock Prediction | Retrieve and preprocess market data for a forecasting experiment. | CLI, Linux, macOS, Windows | The repository is public and licensed under GPL-3.0. |
| Pulse Of Motion | Estimate a video's physical frame rate from its visual dynamics. | Linux | The repository is public. Its README describes a released checkpoint that downloads from Hugging Face on first inference, with a manual download option. The strongest internal model is not fully released; the README directs interested users to contact the project about access. No license information is provided in the repository evidence. |
| Neural Image Assessment | Train the model using AVA images and rating annotations. | Cloud, Linux, macOS, Windows |
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- KittenCN/stock_prediction · Official homepage title · retained in its source language