File size: 1,975 Bytes
3eec7a5 0caf781 3eec7a5 0caf781 3eec7a5 0caf781 3eec7a5 0caf781 3eec7a5 0cefd90 3eec7a5 | 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 74 75 76 | ---
tags:
- generated_from_trainer
datasets:
- Hartunka/processed_wikitext-103-raw-v1-km-100_v2
metrics:
- accuracy
model-index:
- name: tiny_bert_km_100_v2
results:
- task:
name: Masked Language Modeling
type: fill-mask
dataset:
name: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
type: Hartunka/processed_wikitext-103-raw-v1-km-100_v2
metrics:
- name: Accuracy
type: accuracy
value: 0.15307679267672486
---
<!-- 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. -->
# tiny_bert_km_100_v2
This model is a fine-tuned version of [](https://huggingface.co/) on the Hartunka/processed_wikitext-103-raw-v1-km-100_v2 dataset.
It achieves the following results on the evaluation set:
- Loss: 6.9988
- Accuracy: 0.1531
## 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: 0.0001
- train_batch_size: 96
- eval_batch_size: 96
- seed: 10
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10000
- num_epochs: 25
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-------:|:-----:|:---------------:|:--------:|
| 7.1346 | 4.1982 | 10000 | 7.2054 | 0.1477 |
| 6.6576 | 8.3963 | 20000 | 6.9987 | 0.1521 |
| 6.4107 | 12.5945 | 30000 | 7.1060 | 0.1537 |
| 6.2498 | 16.7926 | 40000 | 7.2381 | 0.1518 |
| 6.1573 | 20.9908 | 50000 | 7.1608 | 0.1509 |
### Framework versions
- Transformers 4.40.0
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.19.1
|