metadata

base_model: ai-forever/ruT5-base
tags:
  - generated_from_trainer
model-index:
  - name: ruT5-EmotionNeutralization
    results: []
language:
  - ru
license: mit

results

This model is a fine-tuned version of ai-forever/ruT5-base on an anonymized Emergency calls dataset. It achieves the following results on the evaluation set:

Loss: 0.3848
ParaScore: 0.8265

Model description

More information needed

Intended uses & limitations

Neutralization of emotional speech

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 8

Framework versions

Transformers 4.40.2
Pytorch 2.2.1+cu121
Tokenizers 0.19.1