Tesla V100 32GB to 60 tok/s — every gotcha, measured.
Chat with a local GPT-22M language model in your browser
Chat locally with a 15M‑parameter AI model
Chat with a small AI model directly in your browser