Spaces:
Paused
Paused
update requirements.txt & README.md
Browse files- README.md +75 -7
- requirements.txt +1 -5
README.md
CHANGED
|
@@ -1,14 +1,82 @@
|
|
| 1 |
---
|
| 2 |
title: Dinercall Intent Demo
|
| 3 |
-
emoji:
|
| 4 |
-
colorFrom:
|
| 5 |
-
colorTo:
|
| 6 |
-
sdk:
|
| 7 |
-
sdk_version:
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: apache-2.0
|
| 11 |
-
short_description:
|
| 12 |
---
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
title: Dinercall Intent Demo
|
| 3 |
+
emoji: 🏆
|
| 4 |
+
colorFrom: red
|
| 5 |
+
colorTo: gray
|
| 6 |
+
sdk: gradio
|
| 7 |
+
sdk_version: 5.23.3
|
| 8 |
app_file: app.py
|
| 9 |
pinned: false
|
| 10 |
license: apache-2.0
|
| 11 |
+
short_description: restaurant reservation intent detector
|
| 12 |
---
|
| 13 |
|
| 14 |
+
Absolutely! Here's a full **`README.md`** tailored for your Hugging Face **Streamlit Space**, which features:
|
| 15 |
+
|
| 16 |
+
- Voice recording via mic
|
| 17 |
+
- Transcription using `Jingmiao/whisper-small-zh_tw`
|
| 18 |
+
- Reservation intent classification via ALBERT models
|
| 19 |
+
|
| 20 |
+
---
|
| 21 |
+
|
| 22 |
+
### 📄 `README.md`
|
| 23 |
+
|
| 24 |
+
```markdown
|
| 25 |
+
# 🍽️ 餐廳訂位意圖識別系統 (Mandarin Reservation Intent Classifier)
|
| 26 |
+
|
| 27 |
+
🎙️ 本系統讓使用者可以透過**語音錄音**或**文字輸入**，自動判斷是否具有「訂位意圖」，是語音助理或自動客服前端的理想元件之一。
|
| 28 |
+
|
| 29 |
+
---
|
| 30 |
+
|
| 31 |
+
## 🔍 功能介紹
|
| 32 |
+
|
| 33 |
+
- 🧠 **語音辨識**：使用 fine-tuned Whisper 模型 [`Jingmiao/whisper-small-zh_tw`](https://huggingface.co/Jingmiao/whisper-small-zh_tw) 將語音轉為繁體中文文字。
|
| 34 |
+
- 🤖 **意圖分類**：使用微調的 ALBERT 中文模型判斷輸入是否包含訂位意圖。
|
| 35 |
+
- 📱 **支援手機與桌機**：介面具備良好響應性，適用於各類瀏覽器與行動裝置。
|
| 36 |
+
- 🔊 **瀏覽器錄音**：可直接錄音並即時進行語音辨識與意圖分類。
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
+
|
| 40 |
+
## 🚀 使用方式
|
| 41 |
+
|
| 42 |
+
1. 點擊「▶️ 開始錄音」按鈕開始說話。
|
| 43 |
+
2. 點擊「⏹️ 停止錄音」完成語音輸入。
|
| 44 |
+
3. 系統會自動轉文字，並進行「是否為訂位」意圖判斷。
|
| 45 |
+
4. 或者也可以直接手動輸入文字，再點擊送出按鈕。
|
| 46 |
+
|
| 47 |
+
---
|
| 48 |
+
|
| 49 |
+
## 🤗 使用模型
|
| 50 |
+
|
| 51 |
+
### Whisper 語音辨識：
|
| 52 |
+
- [`Jingmiao/whisper-small-zh_tw`](https://huggingface.co/Jingmiao/whisper-small-zh_tw)
|
| 53 |
+
|
| 54 |
+
### 中文意圖分類模型：
|
| 55 |
+
- [`Luigi/albert-tiny-chinese-dinercall-intent`](https://huggingface.co/Luigi/albert-tiny-chinese-dinercall-intent)
|
| 56 |
+
- [`Luigi/albert-base-chinese-dinercall-intent`](https://huggingface.co/Luigi/albert-base-chinese-dinercall-intent)
|
| 57 |
+
|
| 58 |
+
---
|
| 59 |
+
|
| 60 |
+
## 📦 依賴環境
|
| 61 |
+
|
| 62 |
+
```txt
|
| 63 |
+
streamlit
|
| 64 |
+
transformers>=4.30.0
|
| 65 |
+
torch
|
| 66 |
+
torchaudio
|
| 67 |
+
```
|
| 68 |
+
|
| 69 |
+
---
|
| 70 |
+
|
| 71 |
+
## 🛠️ 開發者備註
|
| 72 |
+
|
| 73 |
+
- 本應用為 Streamlit App，支援 Hugging Face Spaces 部署。
|
| 74 |
+
- 使用 JavaScript 客製化錄音介面，透過 Web Audio API 進行錄音與 POST 回傳。
|
| 75 |
+
- 若需延伸本系統至其他語言或多輪對話，歡迎 fork 本專案進行改造！
|
| 76 |
+
|
| 77 |
+
---
|
| 78 |
+
|
| 79 |
+
© 2024 by [Your Name or Team]. Made with ❤️ using Hugging Face + Streamlit.
|
| 80 |
+
```
|
| 81 |
+
|
| 82 |
+
---
|
requirements.txt
CHANGED
|
@@ -1,8 +1,4 @@
|
|
| 1 |
-
|
| 2 |
transformers>=4.30.0
|
| 3 |
torch
|
| 4 |
-
gradio
|
| 5 |
-
safetensors
|
| 6 |
-
huggingface_hub
|
| 7 |
torchaudio
|
| 8 |
-
gradio>=3.30.0
|
|
|
|
| 1 |
+
streamlit
|
| 2 |
transformers>=4.30.0
|
| 3 |
torch
|
|
|
|
|
|
|
|
|
|
| 4 |
torchaudio
|
|
|