distilbert-imdb-sentiment

Model Card for Model ID

Simply put, this is a sentiment analyzer model. With the evolutions of ai models, a certain question hangs, will ai ever learn what the "SENTIMENT" is. Maybe yes, maybe no, we're uncertain, but this model brings us closer to the reality of AI understanding human sentiments.

Model Details

Model Description

This AI model tells a positive or negative sentiment on a phrase. This model is not a decoder but an encoder classification network model.

Inference Example : Review: 'The cinematography was absolutely breathtaking, but the plot fell completely flat.' Predicted Sentiment: Negative

Review: 'An absolute masterpiece. I haven't enjoyed a movie this much in years!' Predicted Sentiment: Positive

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

• Developed by: Havoc999
• Funded by [kaggle free tier]: [Thank you dear kaggle and gemini for helping me debug errors]
• Shared by [none other than me]
• Model type: [encoder classification network model]
• Language(s) (NLP): [English only]
• License: [The Unlicense]

You can get the onnx version at Havoc999/distilbert-imdb-sentiment-onnx

Training Details

Training Data

The IMDB dataset

Training Procedure

We load a pre-trained DistilBERT model with a sequence classification head configured for 2 classes (Positive / Negative).

Preprocessing [optional]

No preprocessing

Training Hyperparameters

training_args = TrainingArguments( output_dir="./results", learning_rate=2e-5, per_device_train_batch_size=16, per_device_eval_batch_size=16, num_train_epochs=2, weight_decay=0.01, eval_strategy="epoch", save_strategy="epoch", load_best_model_at_end=True, fp16=True, report_to="none" )

Environmental Impact

0.02 kg eq. CO2

Model Card Authors

Havoc999

Model Card Contact

[email protected]