Logger Integration Issues¶

# 1. Login issues
import wandb
wandb.login(key="your_api_key")

# 2. Disable online sync for offline training
trainer = AutoTrainer(
    max_epochs=10,
    logger="wandb",
    logger_kwargs={
        "project": "my-project",
        "offline": True,  # Save logs locally
    },
)

# 3. Resume run
trainer = AutoTrainer(
    logger_kwargs={
        "project": "my-project",
        "id": "run_id",
        "resume": "must",
    },
)

import autotimm as at  # recommended alias
from autotimm import LoggingConfig

# Specify custom log directory
logging_config = LoggingConfig(
    log_dir="./custom_logs",
    log_hyperparameters=True,
)

# View logs
# tensorboard --logdir ./custom_logs

# If port is occupied
# tensorboard --logdir ./custom_logs --port 6007