albert-base-v2-cola / README.md
Alireza1044's picture
add model
7b3d1e4
|
raw
history blame
1.42 kB
---
language:
- en
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- matthews_correlation
model_index:
- name: cola
results:
- task:
name: Text Classification
type: text-classification
dataset:
name: GLUE COLA
type: glue
args: cola
metric:
name: Matthews Correlation
type: matthews_correlation
value: 0.5494768667363472
---
<!-- 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. -->
# cola
This model is a fine-tuned version of [albert-base-v2](https://huggingface.co/albert-base-v2) on the GLUE COLA dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7552
- Matthews Correlation: 0.5495
## 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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4.0
### Training results
### Framework versions
- Transformers 4.9.0
- Pytorch 1.9.0+cu102
- Datasets 1.10.2
- Tokenizers 0.10.3