Merge pull request #128 from stackhpc/fix/gradio-textbox-size

Minor UI improvements and bug fixes

Author: sd109

charts/azimuth-image-analysis/azimuth-ui.schema.yaml

@@ -28,7 +28,5 @@ sortOrder:
   - /azimuth-llm/ui/appSettings/llm_params/frequency_penalty
   - /azimuth-llm/ui/appSettings/llm_params/presence_penalty
   - /azimuth-llm/ui/appSettings/llm_params/top_p
-  # vLLM responds with HTTP 400 BadRequest when top_k is
-  # passed to a vision model (but ollama accepts it)
-  # - /azimuth-llm/ui/appSettings/llm_params/top_k
+  - /azimuth-llm/ui/appSettings/llm_params/top_k
   - /azimuth-llm/api/modelMaxContextLength

charts/azimuth-image-analysis/values.schema.json

@@ -89,6 +89,13 @@
                                             "exclusiveMinimum": 0,
                                             "maximum": 1
                                         },
+                                        "top_k": {
+                                            "type": "integer",
+                                            "title": "LLM Top K",
+                                            "description": "The [top k](https://docs.vllm.ai/en/stable/dev/sampling_params.html) value to use when generating LLM responses (must be an integer).",
+                                            "default": -1,
+                                            "minimum": -1
+                                        },
                                         "presence_penalty": {
                                             "type": "number",
                                             "title": "LLM Presence Penalty",

web-apps/chat/app.py

@@ -185,6 +185,7 @@ def inference_wrapper(*args):
             height="75vh",
             resizable=True,
             sanitize_html=True,
+            autoscroll=False,
             latex_delimiters=[
                 {"left": "$$", "right": "$$", "display": True },
                 {"left": "$", "right": "$", "display": False }

web-apps/image-analysis/app.py

@@ -111,11 +111,14 @@ def analyze_image(image_url, prompt):
 app = gr.Interface(
     fn=analyze_image,
     inputs=[
-        gr.Textbox(label="Image URL"),
-        gr.Textbox(label="Prompt/Question", elem_id="prompt", scale=2),
+        gr.Textbox(label="Image URL", lines=2),
+        gr.Textbox(label="Prompt/Question", elem_id="prompt", scale=2, lines=2),
     ],
-    outputs=[gr.Image(label="Image"), gr.Textbox(label="Results")],
-    flagging_mode='never',
+    outputs=[
+        gr.Image(label="Image"),
+        gr.Textbox(label="Results", lines=2, autoscroll=False),
+    ],
+    flagging_mode="never",
     title=settings.page_title,
     description=settings.page_description,
     examples=[[ex.image_url, ex.prompt] for ex in settings.examples],

web-apps/image-analysis/gradio-client-test.py

@@ -6,13 +6,13 @@
 gradio_host = sys.argv[1]
 
 retries = 60
-for n in range(1, retries+1):
+for n in range(1, retries + 1):
     try:
         client = Client(gradio_host)
         result = client.predict(
             image_url="https://media.licdn.com/dms/image/v2/D4D0BAQHyxNra6_PoUQ/company-logo_200_200/company-logo_200_200/0/1704365018113/stackhpc_ltd_logo?e=1747872000&v=beta&t=Ed3-KZS-sHlg-ne1KC0YjI4Ez7yVvJzWr103nm5eVK0",
-    		prompt="Hi",
-    		api_name="/predict"
+            prompt="Hi",
+            api_name="/predict",
         )
         print(result)
         break