---
library_name: transformers
language:
- hu
license: mit
base_model: openai/whisper-large-v3-turbo
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: whisper-large-v3-turbo-finetuned-hu
  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. -->
# fontos információ, mielött használnád, tesztelnéd

Sajnos úgy tűnik, hogy a Ct2 kvantálás során valami elvész, elromlik a modellben, szinte használhatatlanná válik, az max output tokenek száma drasztikusan leesik, nagyon csonkolja a mondatokat. Még nem tudom hol a határ időben, token számban ahol eklezd csonkolni, de max 10 sec körül vagy inkább alatta.
Natív F32-ben szépen dolgozik ahogy a teszteredmények is mutatják, így viszont elveszti a sebességét, ami az értelme lenne. Lehet újra indítom a finomhangolást nativan Float16-on, hogy lássam azzal a modellel mi történik a kvantálás során. (Az original modell is Float16-ban van)

# whisper-large-v3-turbo-finetuned-hu

This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the custom dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0287
- Wer: 0.0748

## 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: 16
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 2
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step  | Validation Loss | Wer    |
|:-------------:|:------:|:-----:|:---------------:|:------:|
| 0.0574        | 0.1176 | 2000  | 0.0581          | 0.1432 |
| 0.0495        | 0.2352 | 4000  | 0.0517          | 0.1283 |
| 0.0474        | 0.3528 | 6000  | 0.0479          | 0.1184 |
| 0.0454        | 0.4704 | 8000  | 0.0440          | 0.1107 |
| 0.0409        | 0.5880 | 10000 | 0.0416          | 0.1024 |
| 0.0402        | 0.7056 | 12000 | 0.0419          | 0.1045 |
| 0.0377        | 0.8231 | 14000 | 0.0387          | 0.0941 |
| 0.0377        | 0.9407 | 16000 | 0.0371          | 0.0950 |
| 0.0253        | 1.0583 | 18000 | 0.0360          | 0.0899 |
| 0.0244        | 1.1759 | 20000 | 0.0352          | 0.0884 |
| 0.0238        | 1.2935 | 22000 | 0.0342          | 0.0884 |
| 0.023         | 1.4111 | 24000 | 0.0329          | 0.0851 |
| 0.0224        | 1.5287 | 26000 | 0.0320          | 0.0819 |
| 0.0212        | 1.6463 | 28000 | 0.0310          | 0.0805 |
| 0.0196        | 1.7639 | 30000 | 0.0301          | 0.0778 |
| 0.0189        | 1.8815 | 32000 | 0.0292          | 0.0762 |
| 0.0193        | 1.9991 | 34000 | 0.0287          | 0.0748 |


### Framework versions

- Transformers 4.47.0
- Pytorch 2.5.1+cu118
- Datasets 3.1.0
- Tokenizers 0.21.0