Downloads · 30 days
165
1% of all-time downloads
gpt2-imdb-sentiment-classifier
mnoukhov/gpt2-imdb-sentiment-classifier
gpt2-imdb-sentiment-classifier is a text classification model from mnoukhov. Use it when you need a label for a piece of text. It is set up for transformers. The card lists the license as mit.
should probably proofread and complete it, then remove this comment. --
All-time downloads
16.7K
Public
Repo size
2.6 GB
Likes
6
Public
Repo makeup
Click a slice to open those files.
.bin510 MB · 99%
At a glance
- Task
- Text Classification
- Library
- transformers
- License
- mit
- Model type
- gpt2
- Access
- Public
- Created
- Mar 23, 2023
- Updated
- Mar 23, 2023
- SHA
- b3d83413
Try a prompt
Datasets
Model card
From the Hugging Face model README
<!-- 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. -->
gpt2-imdb-sentiment-classifier
This model is a fine-tuned version of gpt2 on the imdb dataset. It achieves the following results on the evaluation set:
- Loss: 0.1703
- Accuracy: 0.9394
Model description
More information needed
Intended uses & limitations
This is comparable to distilbert-imdb and trained with exactly the same script
It achieves slightly lower loss (0.1703 vs 0.1903) and slightly higher accuracy (0.9394 vs 0.928)
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-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: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.1967 | 1.0 | 1563 | 0.1703 | 0.9394 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.13.1+cu117
- Datasets 2.9.0
- Tokenizers 0.12.1
- Task
- Text Classification
- Library
- transformers
- Type
- gpt2
- License
- mit
- Created
- Mar 23, 2023
- Updated
- Mar 23, 2023