What you can do with Kronos

3 use cases

• Forecast future candlesticks from one price history

  Load a pre-trained Kronos model and get a forecast of future OHLCV candles from your own historical data.

  • Kronos
  • Data and analytics
  • Intermediate
  View prompt

  Using the Kronos repo, load NeoQuasar/Kronos-small with its tokenizer, forecast the next 120 candles for <CSV path with open, high, low, close columns and timestamps>, and return the pred_df table. Verify that the forecast index matches the y_timestamp values you passed in and that the lookback is no more than 512 rows.

  Share Save

• Forecast many assets in one batch run

  Use predict_batch to forecast several equal-length price series at once and keep results in input order.

  • Kronos
  • Data and analytics
  • Intermediate
  View prompt

  In the Kronos repo, prepare equal-length candlestick DataFrames for <list of instrument CSV paths>, call predict_batch with pred_len=<number of steps>, and write one forecast CSV per instrument. Verify that the number of returned DataFrames equals the number of input series and that each series has the same lookback length.

  Share Save

• Fine-tune Kronos on your own market dataset

  Adapt the Kronos tokenizer and predictor to your own candlestick data using the Qlib-based finetuning pipeline.

  • Kronos
  • Data and analytics
  • Advanced
  View prompt

  In the Kronos repo, edit finetune/config.py for my local Qlib data at <qlib data path>, run the preprocessing script, then run both torchrun training scripts with <number of GPUs> GPUs. Verify that train_data.pkl, val_data.pkl and test_data.pkl exist in dataset_path before starting training.

  Share Save

How to use Kronos

1. Check you have what it needs: Python 3.10 or newer; pip, and access to install requirements.txt; Internet access to load models from the Hugging Face Hub.
2. Open shiyu-coder/Kronos on GitHub and read the README and license.
3. Run pip install -r requirements.txt in the cloned Kronos folder, using Python 3.10 or newer.
4. Load the tokenizer and model with KronosTokenizer.from_pretrained("NeoQuasar/Kronos-Tokenizer-base") and Kronos.from_pretrained("NeoQuasar/Kronos-small").
5. Create the predictor with KronosPredictor(model, tokenizer, max_context=512).
6. Split your data into a lookback window of 400 rows as x_df, x_timestamp, and a 120-step y_timestamp range, keeping the lookback at or below 512.

These steps come from our top use case, Forecast future candlesticks from one price history, which also has a copy-paste agent prompt.

Kronos alternatives · Best AI repos for Data and analytics

About this repo

• Who it is forQuant analysts and developers who have historical OHLCV candlestick data and want model-based forecasts or a model they can adapt to their own market data.
• Why it is hot
• LanguagePython
• Sourcegithub.com/shiyu-coder/Kronos · Back to Hot repos

Requirements

• Python 3.10 or newer
• pip, and access to install requirements.txt
• Internet access to load models from the Hugging Face Hub
• Historical candlestick data in a pandas DataFrame with open, high, low, close columns
• For fine-tuning only: a local Qlib data directory, a GPU setup that works with torchrun, and optionally a Comet.ml account

Related repos

• Trended earlier DBT Charts dbt Charts is a declarative YAML syntax around SQL for building dashboards, with a dct CLI that validates, previews, renders, and queries board definitions.
• Trended earlier Brandpose BrandPose is a personal project that estimates where shoppers look in a store, using an ordinary fixed security camera and open source AI models (YOLO11, Gaze-LLE, Depth Anything V2).
• Trended earlier Idx A Rust command-line tool and MCP server that gives delayed quotes, fundamentals, screening, news, ownership and broker data for Indonesian stocks (IDX/IHSG) without API keys.