crypto-timeseries-transformer

Overview

Transformer-based regression model for short-term cryptocurrency price forecasting.

Model Architecture

• Encoder-only Transformer
• Continuous regression output

Intended Use

Research, backtesting, feature generation for trading strategies.

Limitations

Market regime shifts can reduce accuracy. Not financial advice.

Example Code

import torch
model = torch.load("pytorch_model.bin")
prediction = model(input_tensor)