File size: 1,755 Bytes
3e446ee 63cd3c3 3e446ee 63cd3c3 3e446ee 63cd3c3 3e446ee 9140015 3e446ee 9140015 63cd3c3 3e446ee |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 |
---
base_model: ad019el/Kabyle_xlsr-finetuned-tamasheq-1
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Kabyle_xlsr-finetuned-tamasheq-1
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. -->
# Kabyle_xlsr-finetuned-tamasheq-1
This model is a fine-tuned version of [ad019el/Kabyle_xlsr-finetuned-tamasheq-1](https://huggingface.co/ad019el/Kabyle_xlsr-finetuned-tamasheq-1) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4590
- Wer: 0.8584
## 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: 3e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 200
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 3.4525 | 15.79 | 300 | 1.4650 | 0.9462 |
| 0.6914 | 31.58 | 600 | 1.2240 | 0.8612 |
| 0.3942 | 47.37 | 900 | 1.3305 | 0.8159 |
| 0.354 | 63.16 | 1200 | 1.3404 | 0.8640 |
| 0.2886 | 78.95 | 1500 | 1.4590 | 0.8584 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|