Merge branch 'main' into feature/rkatakol/ibvs_reverse_proxy

Author: rohitkatakol

.github/CODEOWNERS

@@ -14,13 +14,13 @@ README.md          @xwu2intel
 /manufacturing-ai-suite/weld-porosity/                    @ajagadi1 @sugnanprabhu @rrajore @xwu2intel
 
 # metro-ai
-/metro-ai-suite/                                                 @rrajore @vagheshp @ArokiEdgard @xwu2intel
-/metro-ai-suite/image-based-video-search/                        @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel
+/metro-ai-suite/                                                 @rrajore @vagheshp @ArokiEdgard @xwu2intel @ajagadi1
+/metro-ai-suite/image-based-video-search/                        @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel @ajagadi1
 /metro-ai-suite/interactive-digital-avatar/                      @senhui2intel @Junyu-B @wzq112358 @myqi @xwu2intel
 
-/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/  @vagheshp @tjanczak @xwu2intel
+/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/  @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore
 /metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/   @saratpoluri @sarthakdeva-intel @xwu2intel
-/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/        @vagheshp @tjanczak @xwu2intel
+/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/        @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore
 
 /metro-ai-suite/sensor-fusion-for-traffic-management/            @lijiunderstand @chaofanchen-intel @thegreatchaos @xwu2intel
 /metro-ai-suite/smart-nvr/                                       @bharagha @yogeshmpandey @bhardwaj-nakul @xwu2intel
@@ -52,18 +52,18 @@ README.md          @xwu2intel
 
 
 # documentation content - metro-ai
-/metro-ai-suite/README.md                                       @rrajore @vagheshp @ArokiEdgard @xwu2intel                      @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/image-based-video-search/docs/                  @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel        @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/image-based-video-search/README.md              @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel        @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/README.md                                       @rrajore @vagheshp @ArokiEdgard @xwu2intel @ajagadi1            @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/image-based-video-search/docs/                  @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel @ajagadi1       @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/image-based-video-search/README.md              @rrajore @vagheshp @ArokiEdgard @rohitkatakol @xwu2intel @ajagadi1       @open-edge-platform/open-edge-platform-docs-write
 /metro-ai-suite/interactive-digital-avatar/docs/                @senhui2intel @Junyu-B @wzq112358 @myqi @xwu2intel              @open-edge-platform/open-edge-platform-docs-write
 /metro-ai-suite/interactive-digital-avatar/README.md            @senhui2intel @Junyu-B @wzq112358 @myqi @xwu2intel              @open-edge-platform/open-edge-platform-docs-write
 
-/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/docs/           @vagheshp @tjanczak @xwu2intel                   @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/README.md       @vagheshp @tjanczak @xwu2intel                   @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/docs/            @saratpoluri @sarthakdeva-intel @xwu2intel       @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/README.md        @saratpoluri @sarthakdeva-intel @xwu2intel       @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/docs/                 @vagheshp @tjanczak @xwu2intel                   @open-edge-platform/open-edge-platform-docs-write
-/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/README.md             @vagheshp @tjanczak @xwu2intel                   @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/docs/           @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore                 @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/loitering-detection/README.md       @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore                 @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/docs/            @saratpoluri @sarthakdeva-intel @xwu2intel @ajagadi1 @rrajore       @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/smart-intersection/README.md        @saratpoluri @sarthakdeva-intel @xwu2intel @ajagadi1 @rrajore       @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/docs/                 @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore                  @open-edge-platform/open-edge-platform-docs-write
+/metro-ai-suite/metro-vision-ai-app-recipe/smart-parking/README.md             @vagheshp @tjanczak @xwu2intel @ajagadi1 @rrajore                  @open-edge-platform/open-edge-platform-docs-write
 
 /metro-ai-suite/sensor-fusion-for-traffic-management/docs/      @lijiunderstand @chaofanchen-intel @thegreatchaos @xwu2intel    @open-edge-platform/open-edge-platform-docs-write
 /metro-ai-suite/sensor-fusion-for-traffic-management/README.md  @lijiunderstand @chaofanchen-intel @thegreatchaos @xwu2intel    @open-edge-platform/open-edge-platform-docs-write

education-ai-suite/smart-classroom/README.md

@@ -14,10 +14,10 @@ The goal is to transform raw classroom recordings into concise, structured summa
 ### 💻 System Requirements
 
 - **OS:** Windows 11  
-- **Processor:** Intel® Meteor Lake (with integrated GPU support)  
+- **Processor:** Intel® Core Ultra Series 1 (with integrated GPU support)  
 - **Memory:** 32 GB RAM (minimum recommended)  
 - **Storage:** At least 50 GB free (for models and logs)  
-- **GPU/Accelerator:** Intel® iGPU (Meteor Lake, Arc GPU, or higher) for summarization acceleration  
+- **GPU/Accelerator:** Intel® iGPU (Intel® Core Ultra Series 1, Arc GPU, or higher) for summarization acceleration  
 - **Python:** 3.12 or above  
 - **Node.js:** v18+ (for frontend) 
 ---
@@ -29,7 +29,7 @@ The goal is to transform raw classroom recordings into concise, structured summa
   - Runs on **CPU** (Whisper is CPU-centric)  
 - **FunASR (Paraformer)**  
   - Recommended for **Chinese transcription** (`paraformer-zh`)
-- ✅ Supports transcription of audio files up to 45 minutes
+- ✅ Supports transcription of audio files up to 45 minutes in mp3 and wav formats
 
 #### 🧠 Summarization (LLMs)  
 - **Qwen Models (OpenVINO / IPEX)**  
@@ -50,20 +50,54 @@ The goal is to transform raw classroom recordings into concise, structured summa
 
 **a. Install [FFmpeg](https://ffmpeg.org/download.html)** (required for audio processing):
 
-* On **Windows**:
+- On **Windows**:  
   Download from [https://ffmpeg.org/download.html](https://ffmpeg.org/download.html), and add the `ffmpeg/bin` folder to your system `PATH`.
+---
+
+**Run your shell with admin privileges before starting the application**
 
-**b. Install Python dependencies:**
+**b. Clone Repository:**
 
 ```bash
+  git clone --no-checkout https://github.com/open-edge-platform/edge-ai-suites.git
+  cd edge-ai-suites
+  git sparse-checkout init --cone
+  git sparse-checkout set education-ai-suite
+  git checkout
+  cd education-ai-suite
+```
+---
+
+**c. Install Python dependencies**
+
+It’s recommended to create a **dedicated Python virtual environment** for the base dependencies.
+
+```bash
+python -m venv smartclassroom
+# On Windows:
+smartclassroom\Scripts\activate
+
+python.exe -m pip install --upgrade pip
+pip install --pre --upgrade ipex-llm[xpu_2.6] --extra-index-url https://download.pytorch.org/whl/xpu
 pip install --upgrade -r requirements.txt
 ```
+---
 
-**c. [Optional] Install IPEX-LLM to use IPEX-based LLM model for summarization:**
+
+**d. [Optional] Create Python Venv for Ipex Based Summarizer**  
+If you plan to use IPEX, create a separate virtual environment.
 
 ```bash
+python -m venv smartclassroom_ipex
+# On Windows:
+smartclassroom_ipex\Scripts\activate
+
+python.exe -m pip install --upgrade pip
+pip install --upgrade -r requirements.txt
 pip install --pre --upgrade ipex-llm[xpu_2.6] --extra-index-url https://download.pytorch.org/whl/xpu
 ```
+> 💡 *Use `smartclassroom` if you don’t need IPEX. Use `smartclassroom_ipex` if you want IPEX summarization.*
+
 ---
 ### ⚙️ 2. Default Configuration
 
@@ -92,7 +126,7 @@ asr:
   name: paraformer-zh
 ```
 
-* (Optional) If you want to use IPEX-based summarization, make sure IPEX-LLM is installed and set:
+* (Optional) If you want to use IPEX-based summarization, make sure IPEX-LLM is installed, env for ipex is activated and set following in `config`:
 
 ```bash
 summarizer:
@@ -104,14 +138,17 @@ summarizer:
 ---
 
 ### ✅ 3. **Run the Application**
+Activate the environment before running the application:
 
-Bring Up Backend:
+```bash
+smartclassroom\Scripts\activate  # or smartclassroom_ipex
+```
+Run the backend:
 ```bash
 python main.py
 ```
-**To monitor power usage, run your shell with admin privileges before starting the application.**
 
-Bring Up Frontend:
+- Bring Up Frontend:
 ```bash
 cd ui
 npm install

education-ai-suite/smart-classroom/components/asr/funasr/paraformer.py

@@ -1,6 +1,9 @@
 from components.asr.base_asr import BaseASR
+from utils import ensure_model
+from utils.model_download_helper import get_or_download_model_dir
 from funasr import AutoModel
 
+import os
 import logging
 logger = logging.getLogger(__name__)
 
@@ -10,17 +13,31 @@
     "paraformer-online": "iic/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online",
 }
 
+# use same vad and punc model for different ASR models
+VAD_MODEL = "iic/speech_fsmn_vad_zh-cn-16k-common-pytorch"
+PUNC_MODEL = "iic/punc_ct-transformer_zh-cn-common-vocab272727-pytorch"
+
 class Paraformer(BaseASR):
     def __init__(self, model_name, device="cpu", revision="v2.0.4"):
         if model_name not in FUNASR_MODEL_MAP:
             raise ValueError(f"Invalid ASR model name {model_name}. Supported models are: {list(FUNASR_MODEL_MAP.keys())}")
 
         model_name = FUNASR_MODEL_MAP[model_name]
-        # use same vad and punc model for different ASR models
+        model_dir = ensure_model.get_asr_model_path()
+        model_dir = get_or_download_model_dir(model=model_name, revision=revision, local_dir=model_dir)
+
+        model_dir_parent = os.path.dirname(model_dir) 
+        # download vad model if not exist
+        vad_model_dir = os.path.join(model_dir_parent, VAD_MODEL)
+        vad_model_dir = get_or_download_model_dir(model=VAD_MODEL, revision="v2.0.4", local_dir=vad_model_dir)
+        # download punc model if not exist
+        punc_model_dir = os.path.join(model_dir_parent, PUNC_MODEL)
+        punc_model_dir = get_or_download_model_dir(model=PUNC_MODEL, revision="v2.0.4", local_dir=punc_model_dir)
+
         self.model_name = model_name
-        self.model = AutoModel(model=model_name, model_revision=revision,
-                        vad_model="fsmn-vad", vad_model_revision="v2.0.4",
-                        punc_model="iic/punc_ct-transformer_zh-cn-common-vocab272727-pytorch", punc_model_revision="v2.0.4",
+        self.model = AutoModel(model=model_dir, model_revision=revision,
+                        vad_model=vad_model_dir, vad_model_revision="v2.0.4",
+                        punc_model=punc_model_dir, punc_model_revision="v2.0.4",
                         #   spk_model="cam++", spk_model_revision="v2.0.2",
                         device=device, disable_update=True
                         )

education-ai-suite/smart-classroom/components/llm/ipex/summarizer.py

@@ -4,7 +4,9 @@
 import threading
 from utils.locks import audio_pipeline_lock
 from utils.config_loader import config
+from utils import ensure_model
 import logging
+import os
 logger = logging.getLogger(__name__)
 
 from transformers import TextIteratorStreamer
@@ -50,14 +52,21 @@ def __init__(self, model_name, device="xpu", temperature=0.7):
             logger.info("Loading model in full precision mode.")
             load_in_low_bit = None
 
+        model_dir = ensure_model.get_model_path()
+        local_files_only=False
+        if os.path.exists(model_dir):
+            local_files_only=True
+
         self.model = AutoModelForCausalLM.from_pretrained(
             model_name,
             # load_in_4bit=True,
             load_in_low_bit=load_in_low_bit,
             optimize_model=True,
             trust_remote_code=True,
             use_cache=use_cache,
-            model_hub=model_hub
+            model_hub=model_hub,
+            cache_dir=model_dir,
+            local_files_only=local_files_only
         )
         self.device = device
         self.model = self.model.to(self.device)

education-ai-suite/smart-classroom/docs/toc.rst

@@ -0,0 +1,3 @@
+.. toctree::
+
+   user-guide/index

education-ai-suite/smart-classroom/docs/user-guide/get-started.md

@@ -0,0 +1,123 @@
+# Get Started
+
+This guide walks you through installing dependencies, configuring defaults, and running the application.
+
+## Step 1: Install Dependencies
+
+To install dependencies, do the following:
+
+**a. Install [FFmpeg](https://ffmpeg.org/download.html)** (required for audio processing):
+
+- On **Windows**:  
+  Download from [https://ffmpeg.org/download.html](https://ffmpeg.org/download.html), and add the `ffmpeg/bin` folder to your system `PATH`.
+
+**Run your shell with admin privileges before starting the application**
+
+**b. Clone Repository:**
+
+```bash
+  git clone --no-checkout https://github.com/open-edge-platform/edge-ai-suites.git
+  cd edge-ai-suites
+  git sparse-checkout init --cone
+  git sparse-checkout set education-ai-suite
+  git checkout
+  cd education-ai-suite
+```
+
+**c. Install Python dependencies**
+
+It’s recommended to create a **dedicated Python virtual environment** for the base dependencies.
+
+```bash
+python -m venv smartclassroom
+# On Windows:
+smartclassroom\Scripts\activate
+
+python.exe -m pip install --upgrade pip
+pip install --pre --upgrade ipex-llm[xpu_2.6] --extra-index-url https://download.pytorch.org/whl/xpu
+pip install --upgrade -r requirements.txt
+```
+
+
+**d. [Optional] Create Python Venv for Ipex Based Summarizer**  
+If you plan to use IPEX, create a separate virtual environment.
+
+```bash
+python -m venv smartclassroom_ipex
+# On Windows:
+smartclassroom_ipex\Scripts\activate
+
+python.exe -m pip install --upgrade pip
+pip install --upgrade -r requirements.txt
+pip install --pre --upgrade ipex-llm[xpu_2.6] --extra-index-url https://download.pytorch.org/whl/xpu
+```
+> 💡 *Use `smartclassroom` if you don’t need IPEX. Use `smartclassroom_ipex` if you want IPEX summarization.*
+
+## Step 2: Configure Defaults
+
+The default setup uses Whisper for transcription and OpenVINO Qwen models for summarization. You can customize these in the configuration file.
+
+```bash
+asr:
+  provider: openvino            # Supported: openvino, openai, funasr
+  name: whisper-tiny          # Options: whisper-tiny, whisper-small, paraformer-zh etc.
+  device: CPU                 # Whisper currently supports only CPU
+  temperature: 0.0
+
+summarizer:
+  provider: openvino          # Options: openvino or ipex
+  name: Qwen/Qwen2-7B-Instruct # Examples: Qwen/Qwen1.5-7B-Chat, Qwen/Qwen2-7B-Instruct, Qwen/Qwen2.5-7B-Instruct
+  device: GPU                 # Options: GPU or CPU
+  weight_format: int8         # Supported: fp16, fp32, int4, int8
+  max_new_tokens: 1024        # Maximum tokens to generate in summaries
+```
+### 💡 Tips:
+* For Chinese audio transcription, switch to funASR with Paraformer:
+
+```bash
+asr:
+  provider: funasr
+  name: paraformer-zh
+```
+
+* (Optional) If you are using IPEX-based summarization, make sure IPEX-LLM is installed, env for ipex is activated and set following in `config`:
+
+```bash
+summarizer:
+  provider: ipex
+```
+
+**Important: After updating the configuration, reload the application for changes to take effect.**
+
+## Step 3: Run the Application
+
+Activate the environment before running the application:
+
+```bash
+smartclassroom\Scripts\activate  # or smartclassroom_ipex
+```
+Run the backend:
+```bash
+python main.py
+```
+
+- Bring Up Frontend:
+```bash
+cd ui
+npm install
+npm run dev -- --host 0.0.0.0 --port 5173
+```
+
+## Check Logs
+
+Once the backend starts, you can see the following logs:
+
+```bash
+pipeline initialized
+[INFO] __main__: App started, Starting Server...
+INFO:     Started server process [21616]
+	@@ -92,5 +166,6 @@ INFO:     Application startup complete.
+INFO:     Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)
+```
+
+This means your pipeline server is up and ready to accept requests.

education-ai-suite/smart-classroom/docs/user-guide/index.rst

@@ -0,0 +1,19 @@
+Smart Classroom
+============================================
+The Smart Classroom project is a modular, extensible framework designed to process and summarize educational content using advanced AI models. It supports transcription, summarization, and future capabilities like video understanding and real-time analysis.
+
+The main features are as follows:
+
+- **Audio transcription** with ASR models (e.g., Whisper, Paraformer)  
+-  **Summarization** using powerful LLMs (e.g., Qwen, LLaMA)  
+-  **Plug-and-play architecture** for integrating new ASR and LLM models  
+-  **API-first design** ready for frontend integration  
+-  **Extensible roadmap** for real-time streaming, diarization, translation, and video analysis  
+
+.. toctree::
+   :hidden:
+
+   system-requirements
+   get-started
+   release-notes
+