|
|
--- |
|
|
library_name: transformers |
|
|
license: apache-2.0 |
|
|
base_model: google/t5-efficient-tiny |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
datasets: |
|
|
- generator |
|
|
metrics: |
|
|
- accuracy |
|
|
model-index: |
|
|
- name: sunflower_language_ID_improved |
|
|
results: |
|
|
- task: |
|
|
type: text2text-generation |
|
|
name: Sequence-to-sequence Language Modeling |
|
|
dataset: |
|
|
name: generator |
|
|
type: generator |
|
|
config: default |
|
|
split: train |
|
|
args: default |
|
|
metrics: |
|
|
- type: accuracy |
|
|
value: 0.6293109420681438 |
|
|
name: Accuracy |
|
|
--- |
|
|
|
|
|
<!-- 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. --> |
|
|
|
|
|
# sunflower_language_ID_improved |
|
|
|
|
|
This model is a fine-tuned version of [google/t5-efficient-tiny](https://huggingface.co/google/t5-efficient-tiny) on the generator dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.5044 |
|
|
- Accuracy: 0.6293 |
|
|
- F1 Macro: 0.5576 |
|
|
- F1 Weighted: 0.5783 |
|
|
- Precision Macro: 0.6310 |
|
|
- Recall Macro: 0.6068 |
|
|
|
|
|
## 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: 0.0005 |
|
|
- train_batch_size: 64 |
|
|
- eval_batch_size: 64 |
|
|
- seed: 42 |
|
|
- gradient_accumulation_steps: 2 |
|
|
- total_train_batch_size: 128 |
|
|
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: cosine_with_restarts |
|
|
- lr_scheduler_warmup_steps: 1000 |
|
|
- training_steps: 60000 |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted | Precision Macro | Recall Macro | |
|
|
|:-------------:|:------:|:----:|:---------------:|:--------:|:--------:|:-----------:|:---------------:|:------------:| |
|
|
| 0.8129 | 0.0083 | 500 | 0.9712 | 0.0998 | 0.0544 | 0.0564 | 0.0925 | 0.0963 | |
|
|
| 0.1835 | 0.0167 | 1000 | 0.9716 | 0.2110 | 0.1089 | 0.1089 | 0.1382 | 0.2110 | |
|
|
| 0.1376 | 0.025 | 1500 | 1.1180 | 0.2453 | 0.1461 | 0.1515 | 0.2733 | 0.2365 | |
|
|
| 0.1637 | 0.0333 | 2000 | 0.5585 | 0.4419 | 0.3848 | 0.3991 | 0.4617 | 0.4261 | |
|
|
| 0.1382 | 0.0417 | 2500 | 0.6304 | 0.4811 | 0.4199 | 0.4355 | 0.5272 | 0.4639 | |
|
|
| 0.0589 | 0.05 | 3000 | 0.7011 | 0.4349 | 0.3593 | 0.3726 | 0.4607 | 0.4194 | |
|
|
| 0.1073 | 0.0583 | 3500 | 0.5442 | 0.4991 | 0.4470 | 0.4470 | 0.5804 | 0.4991 | |
|
|
| 0.1461 | 0.0667 | 4000 | 0.4705 | 0.5609 | 0.4802 | 0.4980 | 0.5335 | 0.5408 | |
|
|
| 0.059 | 0.075 | 4500 | 0.5019 | 0.5684 | 0.4987 | 0.4987 | 0.6235 | 0.5684 | |
|
|
| 0.06 | 0.0833 | 5000 | 0.5568 | 0.6106 | 0.5485 | 0.5485 | 0.5973 | 0.6106 | |
|
|
| 0.0617 | 0.0917 | 5500 | 0.4218 | 0.6231 | 0.5450 | 0.5651 | 0.5866 | 0.6008 | |
|
|
| 0.0458 | 0.1 | 6000 | 0.4697 | 0.6276 | 0.5773 | 0.5773 | 0.6620 | 0.6276 | |
|
|
| 0.0646 | 0.1083 | 6500 | 0.4356 | 0.6173 | 0.5432 | 0.5633 | 0.6516 | 0.5952 | |
|
|
| 0.0447 | 0.1167 | 7000 | 0.4705 | 0.6358 | 0.5978 | 0.5978 | 0.6953 | 0.6358 | |
|
|
| 0.0384 | 0.125 | 7500 | 0.4685 | 0.6173 | 0.5600 | 0.5600 | 0.6539 | 0.6173 | |
|
|
| 0.0398 | 0.1333 | 8000 | 0.4796 | 0.6430 | 0.5722 | 0.5933 | 0.6100 | 0.6201 | |
|
|
| 0.0323 | 0.1417 | 8500 | 0.6236 | 0.5705 | 0.5191 | 0.5191 | 0.5960 | 0.5705 | |
|
|
| 0.0344 | 0.15 | 9000 | 0.4619 | 0.6296 | 0.5962 | 0.5962 | 0.7179 | 0.6296 | |
|
|
| 0.0458 | 0.1583 | 9500 | 0.5044 | 0.6293 | 0.5576 | 0.5783 | 0.6310 | 0.6068 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.57.1 |
|
|
- Pytorch 2.8.0+cu126 |
|
|
- Datasets 4.4.1 |
|
|
- Tokenizers 0.22.1 |
|
|
|
