Lightning-AI/pytorch-lightning · error · ValueError

f"`Trainer(barebones=True, log_every_n_steps={log_every_n_st

Error message

f"`Trainer(barebones=True, log_every_n_steps={log_every_n_steps!r})` was passed." " Logging can impact raw speed so it is disabled in barebones mode."

What it means

Trainer was instantiated with barebones=True (speed benchmarking mode) while log_every_n_steps was set to a non-zero value. Barebones mode disables anything that can impact raw speed, including logging, so the constructor rejects incompatible settings immediately instead of silently skewing benchmarks.

Source

Thrown at src/lightning/pytorch/trainer/trainer.py:356

                raise ValueError(
                    f"`Trainer(barebones=True, enable_checkpointing={enable_checkpointing!r})` was passed."
                    " Checkpointing can impact raw speed so it is disabled in barebones mode."
                )
            enable_checkpointing = False
            if logger is not None and logger is not False:
                raise ValueError(
                    f"`Trainer(barebones=True, logger={logger!r})` was passed."
                    " Logging can impact raw speed so it is disabled in barebones mode."
                )
            logger = False
            if enable_progress_bar:
                raise ValueError(
                    f"`Trainer(barebones=True, enable_progress_bar={enable_progress_bar!r})` was passed."
                    " The progress bar can impact raw speed so it is disabled in barebones mode."
                )
            enable_progress_bar = False
            if log_every_n_steps is not None and log_every_n_steps != 0:
                raise ValueError(
                    f"`Trainer(barebones=True, log_every_n_steps={log_every_n_steps!r})` was passed."
                    " Logging can impact raw speed so it is disabled in barebones mode."
                )
            log_every_n_steps = 0
            if enable_model_summary:
                raise ValueError(
                    f"`Trainer(barebones=True, enable_model_summary={enable_model_summary!r})` was passed."
                    " Model summary can impact raw speed so it is disabled in barebones mode."
                )
            enable_model_summary = False
            if num_sanity_val_steps is not None and num_sanity_val_steps != 0:
                raise ValueError(
                    f"`Trainer(barebones=True, num_sanity_val_steps={num_sanity_val_steps!r})` was passed."
                    " Sanity checking can impact raw speed so it is disabled in barebones mode."
                )
            num_sanity_val_steps = 0
            # opt-ins
            if fast_dev_run is not False and fast_dev_run != 0:

View on GitHub (pinned to 9fed5c27d2)

Solutions

Remove log_every_n_steps from the Trainer call when using barebones=True
Set log_every_n_steps=0 or None explicitly
Drop barebones=True if you need logging

Example fix

// before
trainer = Trainer(barebones=True, log_every_n_steps=50)
// after
trainer = Trainer(barebones=True)

Defensive patterns

Strategy: validation

Validate before calling

log_every = cfg.get("log_every_n_steps", 0)
if cfg.get("barebones") and (log_every is not None and log_every != 0):
    del cfg["log_every_n_steps"]
trainer = Trainer(**cfg)

Prevention

Keep a separate minimal benchmark Trainer config instead of toggling barebones on the production one
Remember barebones=True force-disables progress bar, logging, model summary, sanity check and rejects fast_dev_run/detect_anomaly/profiler

When it happens

Trigger: Trainer(barebones=True, log_every_n_steps=50) (any value other than None or 0).

Common situations: Copying a production Trainer config (which usually sets log_every_n_steps=50) and just adding barebones=True for a speed benchmark run.

Related errors

AI-assisted analysis of Lightning-AI/pytorch-lightning@9fed5c27d2 (2026-08-28). Data as JSON: /api/errors/9da461b939da1817. Report an issue: GitHub.