gttbsc_distilbert-freezed-best

Ground truth text based multi-label DAC

Model description

Backbone: DistilBert uncased
Pooling: Self attention
Multi-label classification head: 2 dense layers with two dropouts 0.3 and Tanh activation inbetween

Training and evaluation data

Trained on ground truth slue-phase-2 hvb.
Evaluated on ground truth (GT) and Whisper small transcripts (E2E).

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.00043
• weight decay: 1.5e-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20 (due to small patience stopped at around 3 epoch)

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1