---
library_name: transformers
language:
- zh
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
model-index:
- name: Whisper-Small-squeezeformer-architecture
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper-Small-squeezeformer-architecture

This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Voice_Data_Collection_second_edition dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4571
- Cer: 27.2348

## 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: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 3750
- training_steps: 30000
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Cer      |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 2.9532        | 1.0   | 3750  | 2.9601          | 103.2182 |
| 1.6561        | 2.0   | 7500  | 1.6430          | 85.2058  |
| 0.6747        | 3.0   | 11250 | 0.7233          | 43.9073  |
| 0.4521        | 4.0   | 15000 | 0.5573          | 33.8818  |
| 0.3412        | 5.0   | 18750 | 0.4957          | 29.7393  |
| 0.2109        | 6.0   | 22500 | 0.4640          | 27.6988  |
| 0.1365        | 7.0   | 26250 | 0.4580          | 27.5348  |
| 0.105         | 8.0   | 30000 | 0.4571          | 27.2348  |


### Framework versions

- Transformers 4.44.2
- Pytorch 2.4.0
- Datasets 2.21.0
- Tokenizers 0.19.1