metadata
tags:
- generated_from_trainer
datasets:
- custom_legal_danish_corpus
model-index:
- name: danish-lex-lm-base-mlm
results: []
danish-lex-lm-base-mlm
This model is a fine-tuned version of data/PLMs/danish-lm/danish-lex-lm-base on the custom_legal_danish_corpus dataset. It achieves the following results on the evaluation set:
- Loss: 0.7302
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: 16
- eval_batch_size: 16
- seed: 42
- distributed_type: tpu
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 256
- total_eval_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.05
- training_steps: 500000
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.4648 | 5.36 | 50000 | 1.2920 |
| 1.2165 | 10.72 | 100000 | 1.0625 |
| 1.0952 | 16.07 | 150000 | 0.9611 |
| 1.0233 | 21.43 | 200000 | 0.8931 |
| 0.963 | 26.79 | 250000 | 0.8477 |
| 0.9122 | 32.15 | 300000 | 0.8168 |
| 0.8697 | 37.51 | 350000 | 0.7836 |
| 0.8397 | 42.86 | 400000 | 0.7560 |
| 0.8231 | 48.22 | 450000 | 0.7476 |
| 0.8207 | 53.58 | 500000 | 0.7243 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.12.0+cu102
- Datasets 2.0.0
- Tokenizers 0.12.0