English注册

alakxender/whisper-large-v3-cv17-dv

🤗 Hugging Face 来源automatic-speech-recognitionapache-2.01.5B 参数6.2 GBsafetensors✓ 6 个校验和今天更新
一条命令提交

在你的模型文件夹旁边运行它。它会制作种子、将文件与 Hugging Face 比对，然后提交。你只需开始做种，并粘贴你账户中的密钥。它只读取你的文件，绝不修改。如果愿意，可以先阅读脚本。

curl -fsSL https://pirateface.co/package.sh | bash -s -- --repo alakxender/whisper-large-v3-cv17-dv ./model-folder
需要做种者 →

This model is a fine-tuned version of openai/whisper-large-v3 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4610
  • Wer: 71.0345

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 1e-05
  • train_batch_size: 16
  • eval_batch_size: 36
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.4644 0.9772 300 1.0654 203.9655
0.2384 1.9544 600 0.3342 84.8276
0.1481 2.9316 900 0.2715 78.7931
0.0975 2.9772 1200 0.2635 76.0345
0.0616 3.9544 1500 0.2841 73.1034
0.0399 4.9772 1800 0.3215 72.2414
0.0218 5.9772 2100 0.3881 73.7931
0.046 6.9772 2400 0.2772 74.1379
0.018 7.9544 2700 0.3344 71.3793
0.0067 8.9316 3000 0.3947 71.7241
0.0023 9.9088 3300 0.4246 72.5862
0.0008 10.8860 3600 0.4503 71.7241
0.0003 11.8632 3900 0.4610 71.0345

Framework versions

  • Transformers 4.41.0.dev0
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.0
  • Tokenizers 0.19.1