transformers.js and bumblebee
Bumblebee appears to be a complementary tool to Transformers.js, likely providing a simplified API or additional utilities for running on-device models specifically within the browser environment, leveraging Transformers.js for the underlying inference engine.
About transformers.js
huggingface/transformers.js
State-of-the-art Machine Learning for the web. Run 🤗 Transformers directly in your browser, with no need for a server!
Leverages ONNX Runtime with WASM execution (CPU) or WebGPU (GPU) to run quantized transformer models efficiently in browsers, supporting optional 4-bit and 8-bit quantization. Provides a JavaScript API functionally equivalent to Hugging Face's Python library, enabling seamless conversion of PyTorch, TensorFlow, and JAX models via Optimum. Spans NLP, computer vision, audio, and multimodal tasks across text classification, object detection, speech recognition, and zero-shot learning.
About bumblebee
joshxfi/bumblebee
🐝 Run on-device models directly from your browser via Transformers.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work