WebGPU Embedding Benchmark is a browser-based tool for benchmarking the execution time of BERT-based models using WebGPU and WASM. Users can select different models, batch sizes, and sequence lengths to analyze performance, making it valuable for ML engineers optimizing browser inference.
WebGPU Embedding Benchmark is a LLM eval & observability project. It focuses on evaluating and comparing the performance of BERT-based models in browser environments using WebGPU and WASM. It is built as a B2B product for machine learning engineers. WebGPU Embedding Benchmark costs nothing to use. It ships for the web.
Behind WebGPU Embedding Benchmark is Xenova, and it first shipped in 2024. Key capabilities include model benchmarking, webGPU support, and batch size selection.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Same category — not a similarity match