rotten tomatoes model

Introduction

The rotten-tomatoes-model is a text classification model based on the bert-base-cased architecture, fine-tuned on the Rotten Tomatoes dataset. It predicts the sentiment of movie reviews, labeling them as either negative (LABEL_0) or positive (LABEL_1).

Architecture

The model utilizes the bert-base-cased architecture, which is part of the BERT (Bidirectional Encoder Representations from Transformers) family. It is implemented using the Transformers library, leveraging TensorFlow for model training and deployment.

Training

The model was fine-tuned using the Rotten Tomatoes dataset, consisting of 5,331 positive and 5,331 negative movie reviews. The training process achieved the following results over three epochs:

• Epoch 0: Train Loss: 0.4028, Train Accuracy: 0.8213, Validation Loss: 0.4626, Validation Accuracy: 0.8433
• Epoch 1: Train Loss: 0.1628, Train Accuracy: 0.9390, Validation Loss: 0.3498, Validation Accuracy: 0.8696
• Epoch 2: Train Loss: 0.0386, Train Accuracy: 0.9878, Validation Loss: 0.4790, Validation Accuracy: 0.8621

Guide: Running Locally

1. Setup Environment: Install the necessary libraries.

   pip install transformers==4.18.0 tensorflow==2.8.0 datasets==2.1.0 tokenizers==0.12.1
   

2. Download the Model: Use the Hugging Face Transformers library to load the model.

   from transformers import BertTokenizer, TFBertForSequenceClassification
   
   tokenizer = BertTokenizer.from_pretrained('klin1/rotten-tomatoes-model')
   model = TFBertForSequenceClassification.from_pretrained('klin1/rotten-tomatoes-model')
   

3. Inference: Tokenize and classify your text inputs.

   inputs = tokenizer("Your movie review here", return_tensors="tf")
   outputs = model(inputs)
   predictions = tf.nn.softmax(outputs.logits, axis=-1)
   

Cloud GPUs: For efficient training and inference, consider using cloud-based GPUs like AWS EC2 with NVIDIA GPUs, Google Cloud Platform, or Azure.

License

The rotten-tomatoes-model is licensed under the Apache 2.0 License, allowing for free use and distribution under the terms of the license.