|
|
--- |
|
|
--- |
|
|
base_model: google/gemma-2b |
|
|
library_name: peft |
|
|
pipeline_tag: text-generation |
|
|
tags: |
|
|
- sentiment-analysis |
|
|
- lora |
|
|
- peft |
|
|
- transformers |
|
|
- gemma |
|
|
--- |
|
|
|
|
|
# Sentiment Analyzer Model |
|
|
|
|
|
## Overview |
|
|
|
|
|
This repository contains a **Sentiment Analysis model** fine-tuned using **LoRA (Low-Rank Adaptation)** on top of the **Google Gemma-2B** base model. |
|
|
The model is designed to analyze text input and determine the **sentiment polarity** (positive, negative, or neutral). |
|
|
|
|
|
The model is hosted on Hugging Face and uploaded using the `huggingface_hub` Python API. |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Details |
|
|
|
|
|
### Model Description |
|
|
|
|
|
- **Developed by:** Archana S |
|
|
- **Hugging Face Username:** `archanaachu776` |
|
|
- **Model type:** Text Generation / Sentiment Analysis |
|
|
- **Language(s):** English |
|
|
- **Base Model:** google/gemma-2b |
|
|
- **Fine-tuning Method:** PEFT (LoRA) |
|
|
- **Library:** Transformers + PEFT |
|
|
- **License:** Apache 2.0 |
|
|
- **Finetuned from:** google/gemma-2b |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Sources |
|
|
|
|
|
- **Repository:** https://huggingface.co/archanaachu776/sentiment-analyzer |
|
|
- **Base Model:** https://huggingface.co/google/gemma-2b |
|
|
|
|
|
--- |
|
|
|
|
|
## Intended Uses |
|
|
|
|
|
### Direct Use |
|
|
|
|
|
- Analyze sentiment of user reviews |
|
|
- Customer feedback analysis |
|
|
- Social media sentiment monitoring |
|
|
- Text classification tasks requiring sentiment polarity |
|
|
|
|
|
### Downstream Use |
|
|
|
|
|
- Can be integrated into chatbots |
|
|
- Can be used in recommendation systems |
|
|
- Can be extended for domain-specific sentiment analysis (e.g., product, finance, healthcare) |
|
|
|
|
|
### Out-of-Scope Use |
|
|
|
|
|
- Not suitable for generating medical, legal, or financial advice |
|
|
- Not trained for multilingual sentiment analysis |
|
|
- Not designed for toxicity or hate-speech detection |
|
|
|
|
|
--- |
|
|
|
|
|
## Bias, Risks, and Limitations |
|
|
|
|
|
- The model may reflect biases present in the training data |
|
|
- Performance may vary for informal, slang-heavy, or sarcastic text |
|
|
- Accuracy depends on the domain similarity to training data |
|
|
|
|
|
### Recommendations |
|
|
|
|
|
Users should validate predictions before using them in critical decision-making systems and consider further fine-tuning for domain-specific applications. |
|
|
|
|
|
--- |
|
|
|
|
|
## How to Get Started |
|
|
|
|
|
### Installation |
|
|
|
|
|
```bash |
|
|
pip install transformers peft torch |
|
|
|
|
|
|
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
|
|
|
model_id = "archanaachu776/sentiment-analyzer" |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id) |
|
|
model = AutoModelForCausalLM.from_pretrained(model_id) |
|
|
|
|
|
text = "The product quality is excellent and I am very happy." |
|
|
|
|
|
inputs = tokenizer(text, return_tensors="pt") |
|
|
outputs = model.generate(**inputs, max_new_tokens=50) |
|
|
|
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
|
|
|
