By REI Network | March 6, 2025

Hanabi-1 is the the first in our "Catalog" series, which specializes in Financial Prediction. Catalog will be a series of transformer models designed to serve a variety of different specialized purposes. The majority of these models will be open-sourced, making them freely available to the developer community. For those requiring programmatic integration, our API will provide a seamless way to incorporate these capabilities into existing workflows as well as plugging them to CORE for improved efficiency.

Introducing Hanabi-1: A Transformer model for Financial Market Analysis

While the industry gravitates toward increasingly large models, our research has revealed that financial market prediction benefits from a more specialized, compact architecture. Hanabi-1 demonstrates how targeted design can outperform brute-force approaches in specific domains like financial time series analysis

With 16.4 million parameter model consists of:

• 8 transformer layers with multi-head attention mechanisms

• 384-dimensional hidden states throughout the network

• Multiple specialized predictive pathways for direction, volatility, price change, and spread

• Batch normalization rather than layer normalization for better training dynamics

• Focal loss implementation to address inherent class imbalance

The compact size enables faster inference times and allows us to deploy models at the edge for real-time decision making—critical for high-frequency market environments.

Mathematical Foundations: Functions and Formulas

Positional Encoding

To help the transformer understand sequence ordering, we implement sinusoidal positional encoding:

$PE_{(pos,2i)} = \sin\left(pos \cdot \frac{1}{10000^{2i/d_{model}}}\right)$

$PE_{(pos,2i+1)} = \cos\left(pos \cdot \frac{1}{10000^{2i/d_{model}}}\right)$

Where $pos$ is the position within the sequence and $i$ is the dimension index.

Focal Loss for Direction Prediction

To address the severe class imbalance in financial market direction prediction, we implemented Focal Loss:

$FL(p_t) = -(1 - p_t)^\gamma \log(p_t)$

Where $p_t$ is the model's estimated probability for the correct class and $\gamma$ is the focusing parameter (set to 2.0 in Hanabi-1). This loss function down-weights the contribution of easy examples, allowing the model to focus on harder cases.

Confidence Calibration

A key innovation in Hanabi-1 is our confidence-aware prediction system:

$Confidence = 2 \cdot |p - threshold|$

Where $p$ is the predicted probability and $threshold$ is our calibrated decision boundary (0.5). This allows users to filter predictions based on confidence levels, dramatically improving accuracy in high-confidence scenario.

As shown above, predictions with "High" confidence achieve nearly 100% accuracy, while "Very Low" confidence predictions are barely above random chance.

Training Dynamics and Balanced Validation

Training financial models presents unique challenges, particularly the tendency to collapse toward predicting a single class. Our novel validation scoring function addresses this:

$ValScore = F1 + 0.5 \cdot Accuracy + 0.5 \cdot PR_{balance} - 0.1 \cdot Loss - Balance_{penalty}$

Where $PR_{balance}$ is the precision-recall balance metric:

$PR_{balance} = \frac{\min(Precision, Recall)}{\max(Precision, Recall)}$

And $Balance_{penalty}$ applies severe penalties for extreme prediction distributions:

if precision == 0 or recall == 0:
    # Heavy penalty for predicting all one class
    balance_penalty = 0.5
elif precision < 0.2 or recall < 0.2:
    # Moderate penalty for extreme imbalance
    balance_penalty = 0.3

This scoring function drives the model toward balanced predictions that maintain high accuracy:

The plot above reveals how training progresses through multiple phases, with early fluctuations stabilizing into consistent improvements after epoch 80.

Model Architecture Details

Hanabi-1 employs a specialized architecture with several innovative components:

• Feature differentiation through multiple temporal aggregations:

  • Last hidden state capture (most recent information)

  • Average pooling across the sequence (baseline signal)

  • Attention-weighted aggregation (focused signal)

• Direction pathway with BatchNorm for stable training:

  • Three fully-connected layers with BatchNorm1d

  • LeakyReLU activation (slope 0.1) to prevent dead neurons

  • Xavier initialization with small random bias terms

• Specialized regression pathways:

  • Separate networks for volatility, price change, and spread prediction

  • Reduced complexity compared to the direction pathway

  • Independent optimization focuses training capacity where needed

The model's multi-task design forces the transformer encoder to learn robust representations that generalize across prediction tasks.

Prediction Temporal Distribution

The distribution of predictions over time shows Hanabi-1's ability to generate balanced directional signals across varying market conditions. Green dots represent correct predictions, and red dots are incorrect predictions.

Performance and Future Directions

Current performance metrics:

• Direction accuracy: 73.9%

• F1 score: 0.67

• Balanced predictions: 54.2% positive / 45.8% negative

Hanabi-1 currently operates on two primary configurations:

• 4-hour window model (w4_h1)

• 12-hour window model (w12_h1)

Both predict market movements for the next hour, with the 12-hour window model showing superior performance in more volatile conditions.

Future developments include:

• Extending prediction horizons to 4, 12 and 24 hours

• Implementing adaptive thresholds based on market volatility

• Adding meta-learning approaches for hyperparameter optimization

• Integrating on-chain signals for cross-domain pattern recognition

Conclusion

Hanabi-1 demonstrates that specialized, compact transformers can achieve remarkable results in financial prediction tasks. By focusing on addressing the unique challenges of financial data—class imbalance, temporal dynamics, and confidence calibration—we've created a model that delivers reliable signals even in challenging market conditions.

While the model can still be refined, we found that it’s a robust and important first step towards the definition and creation of even more capable financial models.

Follow the github repo for the current implementation and future upgrades:

GitHub - 0xReisearch/hanabi-1

Contribute to 0xReisearch/hanabi-1 development by creating an account on GitHub.

github.com