|
|
--- |
|
|
base_model: google/gemma-2b |
|
|
library_name: peft |
|
|
pipeline_tag: text-generation |
|
|
tags: |
|
|
- sentiment-analysis |
|
|
- nlp |
|
|
- lora |
|
|
- peft |
|
|
- transformers |
|
|
- business-analytics |
|
|
- social-media-analytics |
|
|
--- |
|
|
|
|
|
# Sentiment Analyzer (LoRA Fine-tuned Gemma-2B) |
|
|
|
|
|
## Model Summary |
|
|
|
|
|
This repository contains a **Sentiment Analysis model** fine-tuned using **LoRA (Low-Rank Adaptation)** on top of **Google’s Gemma-2B** base model. |
|
|
The model is designed for **educational, research, and applied business analytics use cases**, especially sentiment analysis of textual data such as customer feedback and social media content. |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Model Name:** Sentiment Analyzer |
|
|
- **Developed by:** Varun Agrawal |
|
|
- **Hugging Face Username:** `09Vaarun` |
|
|
- **Affiliation:** IIRM Jaipur |
|
|
- **Model Type:** Natural Language Processing (Sentiment Analysis / Text Generation) |
|
|
- **Base Model:** google/gemma-2b |
|
|
- **Fine-tuning Technique:** PEFT (LoRA) |
|
|
- **Language:** English |
|
|
- **License:** Apache 2.0 |
|
|
|
|
|
--- |
|
|
|
|
|
## Intended Use |
|
|
|
|
|
### ✅ Direct Use |
|
|
|
|
|
This model can be used for: |
|
|
|
|
|
- Sentiment analysis of: |
|
|
- Customer reviews |
|
|
- Social media posts |
|
|
- Online feedback forms |
|
|
- Business and marketing text |
|
|
- Academic demonstrations of: |
|
|
- Transformers |
|
|
- Parameter-Efficient Fine-Tuning (PEFT) |
|
|
- LoRA-based adaptation |
|
|
|
|
|
### 🔄 Downstream Use |
|
|
|
|
|
- Social media analytics projects |
|
|
- Business intelligence dashboards |
|
|
- NLP coursework and workshops |
|
|
- Research experiments in sentiment analysis |
|
|
|
|
|
### ❌ Out-of-Scope Use |
|
|
|
|
|
- Medical, legal, or financial decision-making |
|
|
- High-stakes automated systems without human review |
|
|
|
|
|
--- |
|
|
|
|
|
## Bias, Risks, and Limitations |
|
|
|
|
|
- The model may reflect biases present in the training data |
|
|
- Performance may vary across domains and writing styles |
|
|
- Not recommended for critical real-world decisions without further evaluation |
|
|
|
|
|
### Recommendations |
|
|
|
|
|
- Perform domain-specific validation before deployment |
|
|
- Use human oversight for business applications |
|
|
|
|
|
--- |
|
|
|
|
|
## How to Use the Model |
|
|
|
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
|
from peft import PeftModel |
|
|
|
|
|
base_model = "google/gemma-2b" |
|
|
adapter_model = "09Vaarun/sentiment-analyzer" |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(base_model) |
|
|
model = AutoModelForCausalLM.from_pretrained(base_model) |
|
|
model = PeftModel.from_pretrained(model, adapter_model) |
|
|
|
|
|
text = "The service was excellent and the staff was very helpful." |
|
|
inputs = tokenizer(text, return_tensors="pt") |
|
|
|
|
|
outputs = model.generate( |
|
|
**inputs, |
|
|
max_new_tokens=50 |
|
|
) |
|
|
|
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
|
|
|
