tracel-ai/burn · error

Error when sending response through callback channel: {err}

Error message

Error when sending response through callback channel: {err}

What it means

Channel-failure escalation in the async checkpointer thread: after restoring a checkpoint, the worker sends the record back via a oneshot callback; if the receiver was dropped (training loop shut down or interrupted) the send fails and this panic fires — but only when no interrupter is available to stop the loop gracefully. It signals that the caller vanished before reading the restored checkpoint, not a checkpoint corruption.

Source

Thrown at crates/burn-train/src/checkpoint/async_checkpoint.rs:35

struct CheckpointerThread<C, R> {
    checkpointer: C,
    receiver: mpsc::Receiver<Message<R>>,
}

impl<C, R> CheckpointerThread<C, R>
where
    C: Checkpointer<R>,
    R: Checkpoint,
{
    fn run(self) {
        for item in self.receiver.iter() {
            match item {
                Message::Restore(epoch, callback, interrupter) => {
                    let record = self.checkpointer.restore(epoch);
                    callback.send(record).unwrap_or_else(|err| {
                        interrupter.map_or_else(
                            || {
                                panic!(
                                    "Error when sending response through callback channel: {err}"
                                )
                            },
                            |int| int.stop(Some(&err.to_string())),
                        )
                    });
                }
                Message::Save(epoch, state, interrupter) => {
                    self.checkpointer.save(epoch, state).unwrap_or_else(|err| {
                        interrupter.map_or_else(
                            || panic!("Error when saving the state: {err}"),
                            |int| int.stop(Some(&err.to_string())),
                        )
                    });
                }
                Message::Delete(epoch, interrupter) => {
                    self.checkpointer.delete(epoch).unwrap_or_else(|err| {
                        interrupter.map_or_else(

View on GitHub (pinned to d16f7ba2ed)

Solutions

1. Ensure the training loop keeps the callback receiver alive until the restore result is consumed.
2. Check for early exits/panics in the caller thread that drop the channel before receiving.
3. Provide the interrupter so shutdown paths stop the loop instead of panicking on a dead channel.

When it happens

Trigger: Thrown at crates/burn-train/src/checkpoint/async_checkpoint.rs:35 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of tracel-ai/burn@d16f7ba2ed (2026-09-05). Data as JSON: /api/errors/52b46cd8df80e8c3. Report an issue: GitHub.