library_name: transformers datasets: - mitermix/audiosnippets_long_2_8M base_model: - openai/whisper-large-v3-turbo
Trained on 2.6M samples for 7 epochs with linear decreasing learnrate and a warmup of 2000 steps.
