|
|
--- |
|
|
library_name: transformers |
|
|
license: apache-2.0 |
|
|
base_model: distilbert-base-uncased |
|
|
tags: |
|
|
- sentiment analysis |
|
|
- text-classification |
|
|
- distilbert |
|
|
- imdb |
|
|
- transformers |
|
|
metrics: |
|
|
- accuracy |
|
|
model-index: |
|
|
- name: an-imdb-classifier |
|
|
results: [] |
|
|
datasets: |
|
|
- stanfordnlp/imdb |
|
|
--- |
|
|
|
|
|
# an-imdb-classifier |
|
|
|
|
|
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the stanfordnlp.imdb dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.3635 |
|
|
- Accuracy: 0.898 |
|
|
|
|
|
## Model description |
|
|
|
|
|
This model is a fine-tuned version of the distilbert-base-uncased model, trained for sentiment analysis on a subset of the IMDb dataset. |
|
|
It is designed to classify movie reviews as either positive or negative. |
|
|
|
|
|
## Intended uses & limitations |
|
|
|
|
|
This model is intended for use in classifying the sentiment of movie reviews. |
|
|
|
|
|
It can be used for tasks such as: |
|
|
Automatically categorizing movie reviews on websites or platforms. |
|
|
Analyzing the overall sentiment towards a particular movie. |
|
|
Providing feedback to users based on their review sentiment. |
|
|
|
|
|
## Training and evaluation data |
|
|
|
|
|
The model was fine-tuned on a small subset of the IMDb dataset. |
|
|
|
|
|
Training set size: 5000 examples |
|
|
Evaluation set size: 500 examples |
|
|
|
|
|
The dataset contains movie reviews labeled as either positive (label 1) or negative (label 0). |
|
|
The distribution of labels in the training set is approximately equal (2494 negative, 2506 positive). |
|
|
|
|
|
## Training procedure |
|
|
|
|
|
The model was trained using the Hugging Face Trainer on the tokenized IMDb dataset subset, using the preprocess_function to tokenize the text and truncate it. |
|
|
|
|
|
### Training hyperparameters |
|
|
|
|
|
The following hyperparameters were used during training: |
|
|
- learning_rate: 2e-05 |
|
|
- train_batch_size: 16 |
|
|
- eval_batch_size: 16 |
|
|
- seed: 42 |
|
|
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: linear |
|
|
- num_epochs: 3 |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | |
|
|
|:-------------:|:-----:|:----:|:---------------:|:--------:| |
|
|
| No log | 1.0 | 313 | 0.3199 | 0.866 | |
|
|
| 0.2966 | 2.0 | 626 | 0.3023 | 0.89 | |
|
|
| 0.2966 | 3.0 | 939 | 0.3635 | 0.898 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.55.0 |
|
|
- Pytorch 2.6.0+cu124 |
|
|
- Datasets 4.0.0 |
|
|
- Tokenizers 0.21.4 |
