|
--- |
|
language: |
|
- mn |
|
base_model: bayartsogt/mongolian-roberta-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- precision |
|
- recall |
|
- f1 |
|
- accuracy |
|
model-index: |
|
- name: roberta-base-ner-demo-turshilt3 |
|
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. --> |
|
|
|
# roberta-base-ner-demo-turshilt3 |
|
|
|
This model is a fine-tuned version of [bayartsogt/mongolian-roberta-base](https://huggingface.co/bayartsogt/mongolian-roberta-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1163 |
|
- Precision: 0.9235 |
|
- Recall: 0.9346 |
|
- F1: 0.9290 |
|
- Accuracy: 0.9806 |
|
|
|
## 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: 32 |
|
- eval_batch_size: 64 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 15 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy | |
|
|:-------------:|:-------:|:----:|:---------------:|:---------:|:------:|:------:|:--------:| |
|
| 0.78 | 0.9958 | 119 | 0.1344 | 0.7247 | 0.8054 | 0.7629 | 0.9518 | |
|
| 0.1075 | 2.0 | 239 | 0.0992 | 0.8035 | 0.8679 | 0.8344 | 0.9656 | |
|
| 0.0642 | 2.9958 | 358 | 0.0831 | 0.8306 | 0.8849 | 0.8569 | 0.9714 | |
|
| 0.0412 | 4.0 | 478 | 0.0924 | 0.8641 | 0.9022 | 0.8827 | 0.9739 | |
|
| 0.0214 | 4.9958 | 597 | 0.0918 | 0.9064 | 0.9225 | 0.9144 | 0.9778 | |
|
| 0.0143 | 6.0 | 717 | 0.0932 | 0.9189 | 0.9301 | 0.9245 | 0.9801 | |
|
| 0.01 | 6.9958 | 836 | 0.0951 | 0.9199 | 0.9325 | 0.9261 | 0.9803 | |
|
| 0.0074 | 8.0 | 956 | 0.1077 | 0.9207 | 0.9299 | 0.9253 | 0.9795 | |
|
| 0.0053 | 8.9958 | 1075 | 0.1081 | 0.9213 | 0.9329 | 0.9270 | 0.9805 | |
|
| 0.0044 | 10.0 | 1195 | 0.1110 | 0.9223 | 0.9331 | 0.9276 | 0.9806 | |
|
| 0.0037 | 10.9958 | 1314 | 0.1125 | 0.9273 | 0.9362 | 0.9317 | 0.9811 | |
|
| 0.0027 | 12.0 | 1434 | 0.1146 | 0.9250 | 0.9344 | 0.9297 | 0.9807 | |
|
| 0.0023 | 12.9958 | 1553 | 0.1155 | 0.9263 | 0.9361 | 0.9311 | 0.9812 | |
|
| 0.0023 | 14.0 | 1673 | 0.1171 | 0.9242 | 0.9342 | 0.9292 | 0.9805 | |
|
| 0.0021 | 14.9372 | 1785 | 0.1163 | 0.9235 | 0.9346 | 0.9290 | 0.9806 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.40.2 |
|
- Pytorch 2.2.1+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|