Production-ready batch inference for Cohere’s Arabic/English ASR model, with optimized Silero VAD, multi-file GPU batching, subtitles, and optional word-level timestamps.
Run millions of LLM requests through batch APIs (OpenAI, Gemini, Anthropic, vLLM) without silent failures. Plus an open benchmark of batch inference services.
Production-ready batch inference for Cohere’s Arabic/English ASR model, with optimized Silero VAD, multi-file GPU batching, subtitles, and optional word-level timestamps.
