Training
Active DevelopmentTraining-Forge
Memory-driven LLM fine-tuning with OpenVINO inference on Intel hardware
Highlight
3,274 training examples + OpenVINO iGPU inference
🔭
Overview
What Training-Forge is.
Training-Forge is the fine-tuning platform for the Forge ecosystem. It uses Brain-Forge memory data as training sources and runs on Intel hardware — training on CPU with PyTorch, inferencing on iGPU via OpenVINO.
Data Source
Training data is exported from Brain-Forge's 1,419-node memory graph. 5 dataset formats are generated from the graph, producing 3,274 total training examples.