update

README.md

@@ -4,6 +4,7 @@ emoji: 🥇
 colorFrom: green
 colorTo: indigo
 sdk: docker
+disable_embedding: true
 pinned: false
 license: apache-2.0
 tags:

src/execute.py

@@ -111,68 +111,4 @@ def stream_logs(command, jsonl_file=None):
     log_content = []
     for log_line in run_bigcodebench(command):
         log_content.append(log_line)
-        yield "".join(log_content)
-        
-with gr.Blocks() as demo:
-    gr.Markdown("# BigCodeBench Evaluator")
-    
-    with gr.Row():
-        jsonl_file = gr.File(label="Upload JSONL file", file_types=[".jsonl"])
-        split = gr.Dropdown(choices=["complete", "instruct"], label="Split", value="complete")
-        subset = gr.Dropdown(choices=["hard"], label="Subset", value="hard")
-    
-    with gr.Row():
-        parallel = gr.Number(label="Parallel (optional)", precision=0)
-        min_time_limit = gr.Number(label="Min Time Limit", value=1, precision=1)
-        max_as_limit = gr.Number(label="Max AS Limit", value=25*1024, precision=0)
-    
-    with gr.Row():
-        max_data_limit = gr.Number(label="Max Data Limit", value=25*1024, precision=0)
-        max_stack_limit = gr.Number(label="Max Stack Limit", value=10, precision=0)
-        check_gt_only = gr.Checkbox(label="Check GT Only")
-        no_gt = gr.Checkbox(label="No GT")
-    
-    command_output = gr.Textbox(label="Command", value=default_command, interactive=False)
-    with gr.Row():
-        submit_btn = gr.Button("Run Evaluation")
-        download_btn = gr.DownloadButton(label="Download Result")
-    log_output = gr.Textbox(label="Execution Logs", lines=20)
-    
-    input_components = [
-        jsonl_file, split, subset, parallel,
-        min_time_limit, max_as_limit, max_data_limit, max_stack_limit,
-        check_gt_only, no_gt
-    ]
-    
-    for component in input_components:
-        component.change(generate_command, inputs=input_components, outputs=command_output)
-        
-    
-    def start_evaluation(command, jsonl_file, subset, split):
-        extra = subset + "_" if subset != "full" else ""
-        if jsonl_file is not None:
-            result_path = os.path.basename(jsonl_file.name).replace(".jsonl", f"_{extra}eval_results.json")
-        else:
-            result_path = None
-
-        for log in stream_logs(command, jsonl_file):
-            if jsonl_file is not None:
-                yield log, gr.update(value=result_path, label=result_path), gr.update()
-            else:
-                yield log, gr.update(), gr.update()
-        is_running = False
-        result_file = find_result_file()
-        if result_file:
-            return gr.update(label="Evaluation completed. Result file found."), gr.update(value=result_file)
-                    # gr.Button(visible=False)#,
-                    # gr.DownloadButton(label="Download Result", value=result_file, visible=True))
-        else:
-            return gr.update(label="Evaluation completed. No result file found."), gr.update(value=result_path)
-                    # gr.Button("Run Evaluation", visible=True),
-                    # gr.DownloadButton(visible=False))
-    submit_btn.click(start_evaluation,
-                 inputs=[command_output, jsonl_file, subset, split],
-                 outputs=[log_output, download_btn])
-
-demo.queue(max_size=300).launch(share=True, server_name="0.0.0.0", server_port=7860)
-scheduler = BackgroundScheduler()
+        yield "".join(log_content)