Update app.py

app.py

@@ -4,12 +4,11 @@ from huggingface_hub import InferenceClient
 client = InferenceClient()
 
 # Gemma 2
-model_gemma_2 = "google/gemma-2-9b-it"
+model_gemma_2 = "google/gemma-2-27b-it"
 
 def fn_gemma_2(
     prompt,
     history,
-    #input,
     #system_prompt,
     max_tokens,
     temperature,
@@ -26,7 +25,70 @@ def fn_gemma_2(
     #messages.append({"role": "user", "content": prompt})
     #history.append(messages[1])
     
-    #messages = [
+    stream = client.chat.completions.create(
+        model = model_gemma_2,
+        messages = history,
+        max_tokens = max_tokens,
+        temperature = temperature,
+        top_p = top_p,
+        stream = True,
+    )
+    
+    chunks = []
+    for chunk in stream:
+        chunks.append(chunk.choices[0].delta.content or "")
+        yield "".join(chunks)
+
+app_gemma_2 = gr.ChatInterface(
+    fn = fn_gemma_2,
+    type = "messages",
+    additional_inputs = [
+        #gr.Textbox(value="You are a helpful assistant.", label="System Prompt"),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-P"),
+    ],
+    title = "Google Gemma 2",
+    description = model_gemma_2,
+)
+
+# Gemma 3
+model_gemma_3 = "google/gemma-3-27b-it"
+
+def fn_gemma_3(
+    prompt,
+    history,
+    input,
+    #system_prompt,
+    max_tokens,
+    temperature,
+    top_p,
+):
+
+    # Without System Prompt
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {
+                    "type": "text",
+                    "text": prompt
+                },
+                {
+                    "type": "image_url",
+                    "image_url": {
+                        "url": input
+                    }
+                }
+            ]
+        }
+    ]
+    history.append(messages[0])
+
+    # With System Prompt
+    #messages = [{"role": "system", "content": system_prompt}]
+    #history.append(messages[0])
+    #messages.append([
     #    {
     #        "role": "user",
     #        "content": [
@@ -42,11 +104,11 @@ def fn_gemma_2(
     #            }
     #        ]
     #    }
-    #]
-    #history.append(messages[0])
+    #])
+    #history.append(messages[1])
     
     stream = client.chat.completions.create(
-        model = model_gemma_2,
+        model = model_gemma_3,
         messages = history,
         max_tokens = max_tokens,
         temperature = temperature,
@@ -59,21 +121,21 @@ def fn_gemma_2(
         chunks.append(chunk.choices[0].delta.content or "")
         yield "".join(chunks)
 
-app_gemma_2 = gr.ChatInterface(
-    fn = fn_gemma_2,
+app_gemma_3 = gr.ChatInterface(
+    fn = fn_gemma_3,
     type = "messages",
     additional_inputs = [
-        #gr.Textbox(label="Input"),
+        gr.Textbox(label="Input"),
         #gr.Textbox(value="You are a helpful assistant.", label="System Prompt"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-P"),
     ],
-    title = "Google Gemma 2",
-    description = model_gemma_2,
+    title = "Google Gemma 3",
+    description = model_gemma_3,
 )
 
 app = gr.TabbedInterface(
-    [app_gemma_2],
-    ["Gemma 2"]
+    [app_gemma_2, app_gemma_3],
+    ["Gemma 2", "Gemma 3"]
 ).launch()