|
|
--- |
|
|
language: |
|
|
- en |
|
|
|
|
|
pipeline_tag: text-classification |
|
|
|
|
|
tags: |
|
|
- sentiment-analysis |
|
|
- text-classification |
|
|
- opinion-mining |
|
|
- emotion-detection |
|
|
- nlp |
|
|
- natural-language-processing |
|
|
- transformers |
|
|
- peft |
|
|
- lora |
|
|
- adapter |
|
|
- fine-tuning |
|
|
- gemma |
|
|
- gemma-2b |
|
|
- base_model:google/gemma-2b |
|
|
- base_model:adapter:google/gemma-2b |
|
|
|
|
|
library_name: peft |
|
|
|
|
|
base_model: google/gemma-2b |
|
|
|
|
|
license: apache-2.0 |
|
|
|
|
|
model-index: |
|
|
- name: Sentiment Analyzer (LoRA Gemma-2B) |
|
|
results: |
|
|
- task: |
|
|
type: text-classification |
|
|
name: Sentiment Analysis |
|
|
metrics: |
|
|
- type: accuracy |
|
|
value: not-reported |
|
|
--- |
|
|
|
|
|
# Sentiment Analyzer (LoRA Fine-Tuned Gemma-2B) |
|
|
|
|
|
## Model Overview |
|
|
|
|
|
**Sentiment Analyzer** is a **LoRA fine-tuned Gemma-2B transformer model** for **sentiment analysis and text classification** tasks. |
|
|
It uses **PEFT (Parameter-Efficient Fine-Tuning)** to deliver strong performance while keeping memory and compute requirements low. |
|
|
|
|
|
This model is well-suited for: |
|
|
- Sentiment analysis |
|
|
- Opinion mining |
|
|
- Review classification |
|
|
- Emotion-aware text generation |
|
|
- Lightweight NLP deployments |
|
|
|
|
|
--- |
|
|
|
|
|
## Tasks |
|
|
|
|
|
- Text Classification |
|
|
- Sentiment Analysis |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Developed by:** `mysmmurf12` |
|
|
- **Shared by:** `mysmmurf12` |
|
|
- **Model type:** Transformer-based Language Model |
|
|
- **Base model:** `google/gemma-2b` |
|
|
- **Fine-tuning method:** LoRA (Low-Rank Adaptation) |
|
|
- **Library:** PEFT + Transformers |
|
|
- **Language:** English |
|
|
- **License:** Apache 2.0 (inherits from base model) |
|
|
|
|
|
--- |
|
|
|
|
|
## Model Sources |
|
|
|
|
|
- **Hugging Face Repository:** |
|
|
https://huggingface.co/mysmmurf12/sentiment-analyzer |
|
|
|
|
|
- **Base Model:** |
|
|
https://huggingface.co/google/gemma-2b |
|
|
|
|
|
--- |
|
|
|
|
|
## Intended Uses |
|
|
|
|
|
### ✅ Direct Use |
|
|
|
|
|
- Sentiment classification (positive / negative / neutral) |
|
|
- Customer feedback and review analysis |
|
|
- Social media sentiment monitoring |
|
|
- Sentiment-aware chatbots |
|
|
|
|
|
### 🔁 Downstream Use |
|
|
|
|
|
- Integration into RAG pipelines |
|
|
- Domain-specific sentiment fine-tuning |
|
|
- Deployment via APIs, Streamlit apps, or dashboards |
|
|
|
|
|
### 🚫 Out-of-Scope Use |
|
|
|
|
|
- Medical, legal, or financial decision-making |
|
|
- High-risk automated moderation |
|
|
- Multilingual sentiment tasks (English-focused) |
|
|
|
|
|
--- |
|
|
|
|
|
## Bias, Risks, and Limitations |
|
|
|
|
|
- May reflect biases present in training data |
|
|
- Less reliable on sarcasm or ambiguous language |
|
|
- Not evaluated on standardized sentiment benchmarks |
|
|
|
|
|
**Recommendation:** |
|
|
Use human validation for high-impact applications. |
|
|
|
|
|
--- |
|
|
|
|
|
## How to Use the Model |
|
|
|
|
|
### Load with Transformers + PEFT |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
from peft import PeftModel |
|
|
|
|
|
base_model = "google/gemma-2b" |
|
|
adapter_model = "mysmmurf12/sentiment-analyzer" |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(base_model) |
|
|
model = AutoModelForCausalLM.from_pretrained(base_model) |
|
|
|
|
|
model = PeftModel.from_pretrained(model, adapter_model) |
|
|
|
|
|
text = "The product quality is amazing!" |
|
|
inputs = tokenizer(text, return_tensors="pt") |
|
|
|
|
|
outputs = model.generate( |
|
|
**inputs, |
|
|
max_new_tokens=50, |
|
|
temperature=0.7 |
|
|
) |
|
|
|
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
|
|
|
