unslothai/unsloth · error · ValueError

Streaming ChatML-to-Alpaca conversion failed on the first ro

Error message

Streaming ChatML-to-Alpaca conversion failed on the first row: {exc}

What it means

ValueError raised eagerly when convert_chatml_to_alpaca runs on a streaming dataset: the first mapped row is pulled through (next(iter(result))) so per-row failures surface before training instead of mid-iteration. The underlying cause is chained (from exc) — typically _convert hitting a missing conversation column or malformed turns (non-dict messages, missing role/from keys) in row 0. The probe is safe because IterableDataset re-iterates from its source generator.

Source

Thrown at studio/backend/utils/datasets/format_conversion.py:270

        if num_proc is None or type(num_proc) is not int:
            num_proc = dataset_map_num_proc()
        else:
            num_proc = dataset_map_num_proc(num_proc)

        dataset_map_kwargs["num_proc"] = num_proc
        dataset_map_kwargs["desc"] = "Converting ChatML to Alpaca format"

    result = dataset.map(_convert, **dataset_map_kwargs)

    # For streaming, force the first mapped row through now so any
    # column/format errors surface before training begins (not mid-iteration).
    # IterableDataset re-iterates from the generator source, so this is safe.
    if is_iterable:
        try:
            next(iter(result))
        except Exception as exc:
            raise ValueError(
                f"Streaming ChatML-to-Alpaca conversion failed on the first row: {exc}"
            ) from exc

    return result


def convert_alpaca_to_chatml(
    dataset,
    batch_size = 1000,
    num_proc = None,
):
    """
    Convert Alpaca format to ChatML format.

    Output: 'conversations' column with standard 'role'/'content' dicts.
    """
    is_iterable = is_streaming_dataset(dataset)

View on GitHub (pinned to 203007d190)

Solutions

Inspect the chained exception to find the real per-row cause, then fix the data or the column argument
Preview the first row before converting: print(next(iter(dataset))) and confirm the conversation field name/shape
Pass chat_column explicitly rather than relying on the messages/conversations/texts fallback on streaming data

Example fix

# before
result = convert_chatml_to_alpaca(stream_ds)  # raises, chained KeyError('role')

# after
row = next(iter(stream_ds))
print(row.keys(), row['chat'][0])  # confirm column + turn shape
result = convert_chatml_to_alpaca(stream_ds, chat_column='chat')

Defensive patterns

Strategy: try-catch

Validate before calling

def first_row_convertible(stream_ds, chat_column: str | None = None) -> bool:
    row = next(iter(stream_ds), None)
    if row is None:
        return False
    col = chat_column or next(
        (c for c in ("messages", "conversations", "texts") if c in row), None
    )
    if col is None:
        return False
    turns = row[col] or []
    return bool(turns) and isinstance(turns[0], dict)

Try / catch

try:
    result = convert_chatml_to_alpaca(stream_ds, chat_column=col)
except ValueError as e:
    if "failed on the first row" in str(e):
        cause = e.__cause__  # real per-row error (KeyError/TypeError...)
        log_and_surface(cause)
    raise

Prevention

Pull and print the first stream row before converting to verify the column and turn shape
Rely on the chained exception, not the wrapper text, for diagnosis
On streams, pass chat_column explicitly instead of trusting fallback names

When it happens

Trigger: Calling convert_chatml_to_alpaca on an IterableDataset whose rows lack 'messages'/'conversations'/'texts' (or the passed chat_column), or whose first conversation contains turns that are strings/None instead of dicts with role/from keys.

Common situations: Streaming conversions of hub datasets with non-standard column names or mixed-quality first shards; schema drift after upstream producers renamed fields; a chat_column typo that only fails once rows are actually read.

Related errors

AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/9bcacfff5153f9ce. Report an issue: GitHub.