|
--- |
|
license: apache-2.0 |
|
base_model: cl-tohoku/bert-base-japanese-v3 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: jp-speech-classifier |
|
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. --> |
|
|
|
# jp-speech-classifier |
|
|
|
This model is a fine-tuned version of [cl-tohoku/bert-base-japanese-v3](https://huggingface.co/cl-tohoku/bert-base-japanese-v3) on a dataset created from speech records in the Japanese diet. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.1895 |
|
- Accuracy: 0.7053 |
|
|
|
## Model description |
|
|
|
This model classifies Japanese sentences into factual, question, descriptive, opinion based and other sentences. |
|
|
|
## Intended uses & limitations |
|
|
|
This model can be used for any purpose that requires sentence categorization of Japanese sentences. The dataset is fairly small but it gets the job done. |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 2e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 2 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:| |
|
| No log | 1.0 | 72 | 1.1048 | 0.6772 | |
|
| No log | 2.0 | 144 | 1.1895 | 0.7053 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.31.0 |
|
- Pytorch 2.0.1+cu117 |
|
- Datasets 2.13.1 |
|
- Tokenizers 0.13.3 |
|
|
