wav2vec2-large-ami-fine-tuned

This model is a fine-tuned version of facebook/wav2vec2-large-lv60 on the EDINBURGHCSTR/AMI - IHM dataset. It achieves the following results on the evaluation set:

  • Loss: 1.9555
  • Wer: 0.9958

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: 0.00014
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 1000
  • num_epochs: 2.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.5455 0.1565 1000 1.3698 0.8373
1.3019 0.3131 2000 0.7275 0.4146
0.9922 0.4696 3000 0.6047 0.3663
0.5129 0.6262 4000 0.5773 0.3658
0.85 0.7827 5000 0.5387 0.3538
1.4588 0.9393 6000 0.5581 0.3326
0.2646 1.0958 7000 0.5216 0.3294
0.1923 1.2523 8000 0.4975 0.3159
0.2897 1.4089 9000 0.4757 0.3066
0.1536 1.5654 10000 0.4784 0.3066
0.3964 1.7220 11000 0.4899 0.3097
1.1026 1.8785 12000 0.9830 0.8711

Framework versions

  • Transformers 4.42.0.dev0
  • Pytorch 2.3.0a0+gitcd033a1
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
16
Safetensors
Model size
315M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for tz579/wav2vec2-large-ami-fine-tuned

Finetuned
(9)
this model

Dataset used to train tz579/wav2vec2-large-ami-fine-tuned

Evaluation results