Custom Sentiment Analysis Model
This model is a fine-tuned version of distilbert-base-uncased on an IMDB dataset subset for sentiment analysis.
Model Details
Model Description
This model classifies text into two categories: POSITIVE or NEGATIVE. It was developed as part of a custom sentiment analysis project using the Hugging Face transformers and datasets libraries.
- Developed by: Christian DJOMATIN AHO (DJO5555)
- Model type: Transformer-based Sequence Classification
- Language(s) (NLP): English
- License: Apache-2.0
- Finetuned from model: distilbert-base-uncased
Model Sources
Uses
Direct Use
The model can be used directly for sentiment analysis on English text, particularly movie reviews or similar content.
Out-of-Scope Use
The model is not intended for use in production environments requiring high precision, as it was trained on a very small subset of data for demonstration purposes.
Bias, Risks, and Limitations
This model was trained on a small subset (100 samples) of the IMDB dataset, which means its generalization capabilities are limited. It may exhibit biases present in the original IMDB dataset.
Recommendations
Users should be aware that this is a demonstration model. For robust sentiment analysis, a model trained on a larger and more diverse dataset is recommended.
How to Get Started with the Model
Use the code below to get started with the model.
from transformers import pipeline
classifier = pipeline("sentiment-analysis", model="DJO5555/custom-sentiment-analysis")
results = classifier(["I love this AI assistant, it's so helpful!", "This is the worst experience ever."])
for result in results:
print(f"Label: {result['label']}, Score: {result['score']:.4f}")
Training Details
Training Data
The model was trained on a subset of the IMDB dataset.
- Training samples: 100
- Evaluation samples: 50
Training Procedure
Training Hyperparameters
- Training regime: fp32
- Epochs: 1
- Batch size: 8
- Warmup steps: 10
- Weight decay: 0.01
Evaluation
Testing Data, Factors & Metrics
Testing Data
The evaluation was performed on 50 samples from the IMDB test set.
Metrics
The primary metric used for evaluation during training was the training/validation loss.
Results
The model displays basic sentiment recognition capabilities on standard movie review sentences.
Technical Specifications
Model Architecture and Objective
- Architecture: DistilBertForSequenceClassification
- Objective: Single-label Classification (Positive/Negative)
Compute Infrastructure
Software
- Transformers version: 4.57.3
- PyTorch / Hugging Face Trainer API
Model Card Authors
Christian DJOMATIN AHO
Model Card Contact
DJO5555 (Hugging Face Hub)
- Downloads last month
- 30