Relational Transformers

Relational Transformers runs and adapts RT-J over model-ready relational cell tensors. Applications own data retrieval and feature encoding; this library owns the model.

Documentation

• Installation
  • Install with uv
  • Install with pip
  • Install with Conda
  • Install from Source
  • Editable Install
  • Install PyTorch with CUDA support
• Testing
• Quickstart
  • Load RT-J
  • Encoding cells
  • Typed cells
  • Batch prediction
• Computing Predictions
  • Prediction
  • Contextual Cell Embeddings
  • Classification and Regression
  • Batching
  • Ablation
• Prediction
• Relational batches
• Backends
  • PyTorch
  • Triton
  • ONNX
  • Meta
• Efficiency
  • Backend Selection
  • Batch Similar Context Lengths
  • Quantized Models
  • ONNX Export
  • Measure End-to-End Latency
• Ablation
• Custom models
• Pretrained Models
  • Published Models
  • Model Input Contract
  • Classification and Ranking
  • Regression and Forecasting
• Dataset Overview
  • Accepted Input Types
  • Dataset Construction
  • Splitting Relational Data
  • Pre-existing Datasets
• Loss Overview
  • Loss Table
  • Classification
  • Regression and Forecasting
  • Custom Loss Functions
• Quantization
  • FP8
  • Int8
  • Int4
  • Validation
• Training Overview
  • Why Fine-tune?
  • Training Components
  • Model
  • Dataset
  • Loss Function
  • Training Arguments
  • Evaluator
  • Trainer
  • End-to-End Example
• Training overview
• Task-head tuning
• Full-model fine-tuning
• Training Examples
  • Head Tuning
  • Multiclass Classification
  • Full-model Fine-tuning
  • Custom PyTorch Loop
• Model API
  • RelationalTransformer
  • RTJModel
• Batch API
  • RelationalBatch
• Datasets
  • RelationalDataset
• Evaluation
  • BinaryClassificationEvaluator
  • RegressionEvaluator
  • AblationEvaluator
  • SequentialEvaluator
• Losses
  • BinaryClassificationLoss
  • MulticlassClassificationLoss
  • MultilabelClassificationLoss
  • RegressionLoss
• Training API
  • RelationalExample
  • RelationalTrainingArguments
  • RelationalTrainer
  • TaskHead
• ONNX API
  • export_onnx()
  • OnnxBackend