yeyupiaoling/Whisper-Finetune Issues
缺少噪音文件夹问题，dataset/noise
Closed 1whisper-medium微调之后，转换模型失败
Updated 1多卡训练卡死
Updated 1whisper全量微调相关问题
Closed 1Push to Hub Support
Closed 2是否支持mp4 视频
Closed 1微调方言，数据标注为chinese还是新建一个方言的类型呢？
Closed 1建議您整理一下 requirements.txt 及 readme.md
Closed 1NaN or Inf found in input tensor
Closed 2使用large-v2-finetune，高機率出現重複內容…
Closed 2能結合 pyannote 嗎?
Closed 1如何导出pt格式的模型？
Closed 1训练出的模型如何导出npz格式
Closed 1evaluate运行时出现问题
Closed 6ubuntu20.04环境问题
Closed 8transcribe时如何产生时间戳？
Closed 2微调模型保存时需要连接huggingface
Closed 1模型微调后出现乱码
Closed 1请问如何导出onnx
Closed 1请问加速推理可以离线使用吗
Closed 1可以使用initial_prompt做微调吗
Closed 1使用 ct2 轉換後掉辨識率
Closed 2可以单独导出我们微调模型训练出的数据吗
Closed 2可以微调新的语种吗？
Closed 1123
Closed 2如何转换V3版本
Closed 4資料格式的 language 設定
Closed 1微调时的奇怪问题，训练集变大之后，准确度反而下降了
Closed 52卡训练速度比单卡训练快很多
Closed 15正常数据和空数据一起训练的格式
Closed 9如何随机化模型参数，从头开始训练
Closed 2训练稳定后的loss，大家一般都是多少啊
Closed 1一般需要多少数据，可以有不错的效果呢
Closed 1双通道的数据，还是单通道
Closed训练数据里，需要提前把标点符号去掉吗
Closedlora微调large V2版本，需要多大的显存，
Closed 1