Spaces:
Sleeping
Sleeping
Create Report.txt
Browse files- Report.txt +58 -0
Report.txt
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# 📝 Project Report: Multi-Task Language Application with Gradio
|
| 2 |
+
|
| 3 |
+
## ✨ Overview
|
| 4 |
+
|
| 5 |
+
In this project, I designed an interactive web application using **Gradio** with multiple tabs. Each tab showcases a different natural language processing (NLP) capability. The goal was to build a beginner-friendly, user-interactive language tool that demonstrates the power of large language models and voice tools.
|
| 6 |
+
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
## 💡 Tasks & Approaches
|
| 10 |
+
|
| 11 |
+
### 1️⃣ Sentiment Analysis
|
| 12 |
+
- **What it does:** Classifies input text as **positive** or **negative**, showing a confidence score.
|
| 13 |
+
- **Model used:** Default sentiment analysis model (`distilbert-base-uncased-finetuned-sst-2-english`) via Hugging Face.
|
| 14 |
+
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
### 2️⃣ Chatbot
|
| 18 |
+
- **What it does:** Simulates an interactive conversation with the user.
|
| 19 |
+
- **Model used:** `facebook/blenderbot-400M-distill`, a more advanced conversational model for more natural replies.
|
| 20 |
+
|
| 21 |
+
---
|
| 22 |
+
|
| 23 |
+
### 3️⃣ Summarization
|
| 24 |
+
- **What it does:** Generates a concise summary from long input text.
|
| 25 |
+
- **Model used:** `facebook/bart-large-cnn`, a powerful summarization model.
|
| 26 |
+
|
| 27 |
+
---
|
| 28 |
+
|
| 29 |
+
### 4️⃣ Text-to-Speech
|
| 30 |
+
- **What it does:** Converts text into a playable audio file.
|
| 31 |
+
- **Library used:** `gTTS` (Google Text-to-Speech).
|
| 32 |
+
|
| 33 |
+
---
|
| 34 |
+
|
| 35 |
+
## ⚙️ Technologies & Libraries
|
| 36 |
+
|
| 37 |
+
- **Gradio:** For creating the web-based interactive interface with tabs.
|
| 38 |
+
- **Transformers (Hugging Face):** For accessing pre-trained NLP models.
|
| 39 |
+
- **Torch:** Backend framework for the models.
|
| 40 |
+
- **gTTS:** For converting text to speech.
|
| 41 |
+
|
| 42 |
+
---
|
| 43 |
+
|
| 44 |
+
## 🚧 Challenges Faced
|
| 45 |
+
|
| 46 |
+
- Understanding how to integrate different models into one multi-tab interface.
|
| 47 |
+
- Managing various input/output types (text, audio) in Gradio.
|
| 48 |
+
- Using larger conversational models and handling token limits.
|
| 49 |
+
|
| 50 |
+
---
|
| 51 |
+
|
| 52 |
+
## ✅ Conclusion
|
| 53 |
+
|
| 54 |
+
This project helped me explore Gradio's Blocks system and integrate multiple language tasks into one easy-to-use interface. It demonstrates practical applications of sentiment analysis, conversational AI, summarization, and text-to-speech — all in a single, accessible web app.
|
| 55 |
+
|
| 56 |
+
---
|
| 57 |
+
|
| 58 |
+
### 📄 Requirements
|