imdb-sentiment-model

This model is a fine-tuned version of distilbert-base-uncased on the imdb dataset. It achieves the following results on the evaluation set:

• Loss: 0.3340
• Accuracy: 0.896

Model description

This model takes the distilbert-base-uncased model and performs fine-tuning on it using the imdb dataset.

Intended uses & limitations

Mostly to be used to perform sentiment analysis on movie reviews.

How to use

Here is an example of using the model to predict movie review sentiment in PyTorch:

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

checkpoint = "thundarain018/imdb-sentiment-model"
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForSequenceClassification.from_pretrained(checkpoint)

input_text = "I didn't enjoy this movie at all!"

encoded_input = tokenizer(input_text, return_tensors='pt')
output = model(**encoded_input)
predictions = torch.nn.functional.softmax(output.logits, dim=-1)

print(f"Positive score: {predictions[0][1].item():.2%}")
print(f"Negative score: {predictions[0][0].item():.2%}")

Positive score: 2.48%
Negative score: 97.52%

Alternatively you can use this model with a pipeline directly as well.

Training data

The fine tuning dataset used was the Stanford imdb dataset

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	0.3541	0.858
0.2952	2.0	626	0.3340	0.896
0.2952	3.0	939	0.3544	0.905

Framework versions

• Transformers 4.50.2
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1