File size: 2,041 Bytes
a9ceb3d |
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 68 69 70 71 72 73 |
---
base_model: lnxdx/B2_1000_1e-5_hp-mehrdad
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: C1_1000_1e-5_mehrdad
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. -->
# C1_1000_1e-5_mehrdad
This model is a fine-tuned version of [lnxdx/B2_1000_1e-5_hp-mehrdad](https://huggingface.co/lnxdx/B2_1000_1e-5_hp-mehrdad) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6620
- Wer: 0.3151
## 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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 1000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 0.7062 | 0.62 | 100 | 0.6594 | 0.3160 |
| 0.7111 | 1.25 | 200 | 0.6545 | 0.3178 |
| 0.7094 | 1.88 | 300 | 0.6425 | 0.3218 |
| 0.649 | 2.5 | 400 | 0.6549 | 0.3186 |
| 0.6436 | 3.12 | 500 | 0.6657 | 0.3154 |
| 0.6541 | 3.75 | 600 | 0.6530 | 0.3172 |
| 0.6317 | 4.38 | 700 | 0.6601 | 0.3151 |
| 0.6405 | 5.0 | 800 | 0.6589 | 0.3180 |
| 0.6284 | 5.62 | 900 | 0.6641 | 0.3143 |
| 0.6287 | 6.25 | 1000 | 0.6620 | 0.3151 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
|