You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
QAI AppBuilder provides examples for building Vision Language Model (VLM) applications that combine image and video understanding with natural language processing.
216
-
217
-
#### Running VLM Examples
218
-
219
-
Navigate to the VLM samples directory and run the demo:
220
-
221
-
```bash
222
-
cd samples/linux/python/qwen2_vl
223
-
python demo_app.py <model_path>
224
-
```
225
-
226
-
**Parameters:**
227
-
-`<model_path>`: Path to the directory containing the Qwen2-VL QNN model files
228
-
229
-
**VLM Capabilities:**
230
-
- Image/Video/Web camera input processing
231
-
- Multimodal inference combining vision and language
232
-
- Natural language output generation
233
-
- Qwen2-VL-2B-Instruct model support
234
-
235
-
**For detailed setup and usage instructions, see:**[Qwen2-VL Demo (Linux Python)](./python/qwen2_vl/README.md)
236
-
237
213
238
214
## Advanced Application Examples
239
215
@@ -243,7 +219,7 @@ In addition to the computer vision samples above, QAI AppBuilder provides advanc
243
219
244
220
1. **Chat Application with OpenAI-Compatible API**: Build conversational chat WebUI applications using OpenAI-compatible APIs for seamless integration.
245
221
246
-
2.**LangFlow Low-Code Framework**: Deploy and run the LangFlow low-code framework on IQ9075 for rapid AI application development.
222
+
2. **Vision Language Model Examples**: Example for building Vision Language Model (VLM) applications that combine image and video understanding with natural language processing..
247
223
248
224
### Advanced Example 1: Chat Application with OpenAI-Compatible API
249
225
@@ -266,26 +242,30 @@ Build a chat application powered by large language models (LLMs) using OpenAI-co
266
242
3.**Access the WebUI:**
267
243
Open your browser and navigate to `http://localhost:7860` (default port)
268
244
269
-
### Advanced Example 2: LangFlow Low-Code Framework
270
245
271
-
Deploy LangFlow, a visual low-code framework for building AI applications with drag-and-drop components.
246
+
### Advanced Example 2: Vision Language Model Examples
272
247
273
-
**Steps to Run:**
248
+
QAI AppBuilder provides examples for building Vision Language Model (VLM) applications that combine image and video understanding with natural language processing.
274
249
275
-
1.**Start the LLM Service:**
276
-
```bash
277
-
cd tools/launcher_linux
278
-
bash ./4.Start_GenieAPIService.sh
279
-
```
250
+
#### Running VLM Examples
280
251
281
-
2.**Install and Launch LangFlow:**
282
-
```bash
283
-
bash ./5.Install_LangFlow.sh
284
-
bash ./6.Start_LangFlow.sh
285
-
```
252
+
Navigate to the VLM samples directory and run the demo:
286
253
287
-
Once started, access the LangFlow web interface to design and deploy your AI workflows visually.
254
+
```bash
255
+
cd samples/linux/python/qwen2_vl
256
+
python demo_app.py <model_path>
257
+
```
258
+
259
+
**Parameters:**
260
+
-`<model_path>`: Path to the directory containing the Qwen2-VL QNN model files
288
261
262
+
**VLM Capabilities:**
263
+
- Image/Video/Web camera input processing
264
+
- Multimodal inference combining vision and language
265
+
- Natural language output generation
266
+
- Qwen2-VL-2B-Instruct model support
267
+
268
+
**For detailed setup and usage instructions, see:**[Qwen2-VL Demo (Linux Python)](./python/qwen2_vl/README.md)
0 commit comments