|
--- |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: first |
|
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. --> |
|
|
|
# first |
|
|
|
This model is a fine-tuned version of [longformer-gottbert-base-8192-aw512-](https://huggingface.co/longformer-8192-aw512-gottbert-base) on the a 500 million token subset of the german parts of the OSCAR dataset. |
|
It achieves the following results on the custom evaluation set: |
|
- Loss: 1.4981 |
|
|
|
## Model description |
|
|
|
The weights of the model are initialized from the german version of Roberta [gottbert-base](https://huggingface.co/uklfr/gottbert-base). |
|
The local attention windows have a fixed size of 512 tokens across all layers. |
|
The maximum sequence length is 8192. |
|
|
|
## Intended uses & limitations |
|
|
|
Longformer models enable processing long texts using a mixture of local attention on each subword token and task specific global attention on a subset of the tokens. |
|
|
|
## Training and evaluation data |
|
|
|
The [OSCAR](https://oscar-corpus.com) dataset is freely avaible corpus of filtered web texts from the Common Crawl in various languages. We used the 2017 version of the dataset. |
|
|
|
## Training procedure |
|
The model was trained with masked language modeling for 3 epochs on a customly created 500 million tokens subset of the german proportion of the [OSCAR](https://oscar-corpus.com) dataset. |
|
It was validated using 5% of the original subset. |
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 3e-05 |
|
- train_batch_size: 2 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 8 |
|
- 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 |
|
- num_epochs: 3.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:-----:|:---------------:| |
|
| 2.5636 | 0.1 | 500 | 2.2399 | |
|
| 2.0426 | 0.2 | 1000 | 1.8841 | |
|
| 1.9653 | 0.3 | 1500 | 1.7807 | |
|
| 1.9422 | 0.4 | 2000 | 1.7206 | |
|
| 1.9323 | 0.49 | 2500 | 1.6800 | |
|
| 1.7587 | 0.59 | 3000 | 1.6507 | |
|
| 1.7239 | 0.69 | 3500 | 1.6316 | |
|
| 1.7452 | 0.79 | 4000 | 1.6137 | |
|
| 1.7415 | 0.89 | 4500 | 1.5983 | |
|
| 1.7733 | 0.99 | 5000 | 1.5830 | |
|
| 1.7656 | 1.09 | 5500 | 1.5735 | |
|
| 1.6543 | 1.19 | 6000 | 1.5643 | |
|
| 1.7131 | 1.28 | 6500 | 1.5546 | |
|
| 1.6456 | 1.38 | 7000 | 1.5503 | |
|
| 1.716 | 1.48 | 7500 | 1.5422 | |
|
| 1.806 | 1.58 | 8000 | 1.5377 | |
|
| 1.8407 | 1.68 | 8500 | 1.5327 | |
|
| 1.6371 | 1.78 | 9000 | 1.5278 | |
|
| 1.6453 | 1.88 | 9500 | 1.5231 | |
|
| 1.7754 | 1.98 | 10000 | 1.5214 | |
|
| 1.7695 | 2.08 | 10500 | 1.5165 | |
|
| 1.7109 | 2.17 | 11000 | 1.5138 | |
|
| 1.6992 | 2.27 | 11500 | 1.5107 | |
|
| 1.6707 | 2.37 | 12000 | 1.5097 | |
|
| 1.6835 | 2.47 | 12500 | 1.5040 | |
|
| 1.7171 | 2.57 | 13000 | 1.5041 | |
|
| 1.7257 | 2.67 | 13500 | 1.4990 | |
|
| 1.6287 | 2.77 | 14000 | 1.5017 | |
|
| 1.7737 | 2.87 | 14500 | 1.4983 | |
|
| 1.4002 | 2.96 | 15000 | 1.4992 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.15.0 |
|
- Pytorch 1.10.1+cu113 |
|
- Datasets 1.17.0 |
|
- Tokenizers 0.10.3 |
|
|
