|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: mnli_MULTI |
|
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. --> |
|
|
|
# mnli_MULTI |
|
|
|
This model is a fine-tuned version of [WillHeld/roberta-base-mnli](https://huggingface.co/WillHeld/roberta-base-mnli) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5379 |
|
- Acc: 0.8407 |
|
|
|
## 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: 2e-05 |
|
- train_batch_size: 32 |
|
- eval_batch_size: 32 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Acc | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:| |
|
| 0.4798 | 0.17 | 2000 | 0.4797 | 0.8220 | |
|
| 0.4483 | 0.33 | 4000 | 0.4576 | 0.8257 | |
|
| 0.4282 | 0.5 | 6000 | 0.4628 | 0.8298 | |
|
| 0.4333 | 0.67 | 8000 | 0.4426 | 0.8311 | |
|
| 0.4251 | 0.83 | 10000 | 0.4449 | 0.8333 | |
|
| 0.4154 | 1.0 | 12000 | 0.4518 | 0.8376 | |
|
| 0.3064 | 1.17 | 14000 | 0.4845 | 0.8350 | |
|
| 0.3072 | 1.33 | 16000 | 0.4803 | 0.8383 | |
|
| 0.3144 | 1.5 | 18000 | 0.4824 | 0.8322 | |
|
| 0.3025 | 1.66 | 20000 | 0.4634 | 0.8391 | |
|
| 0.3063 | 1.83 | 22000 | 0.4625 | 0.8419 | |
|
| 0.304 | 2.0 | 24000 | 0.4637 | 0.8401 | |
|
| 0.2232 | 2.16 | 26000 | 0.5330 | 0.8434 | |
|
| 0.2215 | 2.33 | 28000 | 0.5338 | 0.8374 | |
|
| 0.2181 | 2.5 | 30000 | 0.5366 | 0.8393 | |
|
| 0.216 | 2.66 | 32000 | 0.5333 | 0.8425 | |
|
| 0.2118 | 2.83 | 34000 | 0.5293 | 0.8414 | |
|
| 0.2128 | 3.0 | 36000 | 0.5379 | 0.8407 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.24.0 |
|
- Pytorch 1.13.0+cu117 |
|
- Datasets 2.7.1 |
|
- Tokenizers 0.12.1 |
|
|
