Update ui_components.py

ui_components.py

@@ -139,17 +139,12 @@ class UIComponents:
             with gr.Row():
                 self.temperature_slider = gr.Slider(minimum=0.0, maximum=2.0, value=0.3, step=0.01, label="Temperature")
                 self.top_p_slider = gr.Slider(minimum=0.0, maximum=1.0, value=1.0, step=0.01, label="Top-p")
-                self.top_logprobs = gr.Slider(minimum=0, maximum=5, value=0, step=1, label="Top logprobs")
             with gr.Row():
                 self.max_tokens_box = gr.Number(value=8192, precision=0, label="Max tokens")
-                self.seed_box = gr.Number(value=None, precision=0, label="Seed")
+                self.seed_box = gr.Number(value=None, precision=0, label="Seed (-1 = random)")
             with gr.Row():
                 self.frequency_penalty = gr.Slider(minimum=-2.0, maximum=2.0, value=0.0, step=0.01, label="Frequency penalty")
                 self.presence_penalty = gr.Slider(minimum=-2.0, maximum=2.0, value=0.0, step=0.01, label="Presence penalty")
-            with gr.Row():
-                self.logprobs_chk = gr.Checkbox(value=False, label="Return logprobs")
-                self.stream_chk = gr.Checkbox(value=False, label="Stream responses (not yet supported in UI)")
-                self.stream_include_usage = gr.Checkbox(value=False, label="Stream include usage")
             self.stop_sequences = gr.Textbox(label="Stop sequences (comma-separated)", placeholder="e.g. \n\n, User:")
 
             # Reasoning effort (GPT-OSS)
@@ -176,31 +171,24 @@ class UIComponents:
             self.response_format.change(_on_response_format_change, inputs=[self.response_format], outputs=[self.json_schema_group])
 
             def update_generation_params(
-                temperature, top_p, top_logprobs, max_tokens, seed,
-                frequency_penalty, presence_penalty, logprobs, stream,
+                temperature, top_p, max_tokens, seed,
+                frequency_penalty, presence_penalty,
                 stop_sequences, reasoning_effort, response_format,
                 json_schema_name, json_schema_description, json_schema_editor, json_schema_strict,
-                tool_choice, tool_function_name, tool_prompt, tools_json, stream_include_usage
+                tool_choice, tool_function_name, tool_prompt, tools_json
             ):
                 params = {
                     "temperature": float(temperature) if temperature is not None else None,
                     "top_p": float(top_p) if top_p is not None else None,
-                    "top_logprobs": int(top_logprobs) if top_logprobs else None,
                     "max_tokens": int(max_tokens) if max_tokens else None,
-                    "seed": int(seed) if seed not in (None, "") else None,
+                    "seed": (None if seed in (None, "", -1) else int(seed)),
                     "frequency_penalty": float(frequency_penalty) if frequency_penalty is not None else None,
                     "presence_penalty": float(presence_penalty) if presence_penalty is not None else None,
-                    "logprobs": bool(logprobs),
-                    # Prevent enabling streaming until UI supports it end-to-end
-                    "stream": False,
                     # stop: list[str]
                     "stop": [s.strip() for s in stop_sequences.split(",") if s.strip()] if stop_sequences else None,
                     # GPT-OSS specific control stored separately; mcp_client merges it
                     "reasoning_effort": reasoning_effort,
                 }
-                # stream options
-                if stream_include_usage:
-                    params["stream_options"] = {"include_usage": True}
 
                 # response_format
                 if response_format == "json_object":
@@ -253,24 +241,22 @@ class UIComponents:
 
             # Wire updates on change
             for comp in [
-                self.temperature_slider, self.top_p_slider, self.top_logprobs,
+                self.temperature_slider, self.top_p_slider,
                 self.max_tokens_box, self.seed_box, self.frequency_penalty,
-                self.presence_penalty, self.logprobs_chk, self.stream_chk,
+                self.presence_penalty,
                 self.stop_sequences, self.reasoning_effort, self.response_format,
                 self.json_schema_name, self.json_schema_description, self.json_schema_editor, self.json_schema_strict,
-                self.tool_choice, self.tool_function_name, self.tool_prompt, self.tools_json,
-                self.stream_include_usage
+                self.tool_choice, self.tool_function_name, self.tool_prompt, self.tools_json
             ]:
                 comp.change(
                     update_generation_params,
                     inputs=[
-                        self.temperature_slider, self.top_p_slider, self.top_logprobs,
+                        self.temperature_slider, self.top_p_slider,
                         self.max_tokens_box, self.seed_box, self.frequency_penalty,
-                        self.presence_penalty, self.logprobs_chk, self.stream_chk,
+                        self.presence_penalty,
                         self.stop_sequences, self.reasoning_effort, self.response_format,
                         self.json_schema_name, self.json_schema_description, self.json_schema_editor, self.json_schema_strict,
-                        self.tool_choice, self.tool_function_name, self.tool_prompt, self.tools_json,
-                        self.stream_include_usage
+                        self.tool_choice, self.tool_function_name, self.tool_prompt, self.tools_json
                     ],
                     outputs=[self.gen_param_status]
                 )