I just used 3,000 GPU-hours to test all 9 new OpenAI Whisper speech recognition models, two Talon acoustic models, and NVIDIA Nemo large and xlarge models. Whisper has a peculiar failure case. Here's what I think.

I just used 3,000 GPU-hours to test all 9 new OpenAI Whisper speech recognition models, two Talon acoustic models, and NVIDIA Nemo large and xlarge models. Whisper has a peculiar failure case. Here's what I think.