tensorboard_launch_with_rust_data_server_backend.py

1import os
2import subprocess
3
4# tensorboard-data-server is used automatically by TensorBoard 
5# when installed in the same environment. 
6# There is no direct Python API for end-user scripting.
7
8def start_tensorboard(logdir):
9    # Ensure the log directory exists
10    if not os.path.exists(logdir):
11        os.makedirs(logdir)
12        print(f"Created log directory: {logdir}")
13
14    print("Launching TensorBoard...")
15    print("Note: tensorboard-data-server will be used automatically if installed.")
16    
17    # Launch TensorBoard as a subprocess
18    # The data server handles the reading of tfevents files in the background
19    subprocess.run(["tensorboard", "--logdir", logdir])
20
21if __name__ == "__main__":
22    # Example directory
23    LOG_DIR = "./logs"
24    start_tensorboard(LOG_DIR)