unslothai/unsloth · error · ValueError

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

Error message

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

What it means

ValueError raised eagerly when convert_alpaca_to_chatml runs on a streaming dataset: the first mapped row is forced through so errors surface at setup time, not mid-training. The chained exception (from exc) holds the real cause — typically the Alpaca source rows missing the instruction/output keys that _convert reads when building conversations. Safe on streams because IterableDataset re-iterates from its generator.

Source

Thrown at studio/backend/utils/datasets/format_conversion.py:336

        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 Alpaca to ChatML 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 Alpaca-to-ChatML conversion failed on the first row: {exc}"
            ) from exc

    return result


def _format_eta(seconds):
    """Format seconds into a human-readable ETA string."""
    if seconds < 60:
        return f"{seconds:.0f}s"
    elif seconds < 3600:
        m, s = divmod(int(seconds), 60)
        return f"{m}m {s}s"
    else:
        h, remainder = divmod(int(seconds), 3600)
        m, _ = divmod(remainder, 60)
        return f"{h}h {m}m"

View on GitHub (pinned to 203007d190)

Solutions

Read the chained exc to identify the missing field or type error, then fix the source data or key mapping
Confirm the dataset really is Alpaca: print(next(iter(dataset)).keys()) should show instruction/output-style fields
Rename variant keys first (e.g. rename_column('prompt','instruction')) or pre-map rows to the expected Alpaca schema before converting

Example fix

# before
result = convert_alpaca_to_chatml(stream_ds)  # rows use prompt/response -> chained error

# after
stream_ds = stream_ds.rename_column('prompt', 'instruction').rename_column('response', 'output')
result = convert_alpaca_to_chatml(stream_ds)

Defensive patterns

Strategy: try-catch

Validate before calling

def first_row_is_alpaca(stream_ds) -> bool:
    row = next(iter(stream_ds), None)
    if row is None:
        return False
    return "instruction" in row and "output" in row

Type guard

def is_alpaca_row(row: dict) -> bool:
    return isinstance(row, dict) and "instruction" in row and "output" in row

Try / catch

try:
    result = convert_alpaca_to_chatml(stream_ds)
except ValueError as e:
    if "failed on the first row" in str(e):
        cause = e.__cause__  # e.g. KeyError('instruction') on prompt/response schemas
        log_and_surface(cause)
    raise

Prevention

Verify the stream really is Alpaca (instruction/output fields) before converting
Rename variant keys (prompt/response -> instruction/output) up front
Diagnose via the chained __cause__; the wrapper only tells you it failed on row 0

When it happens

Trigger: Calling convert_alpaca_to_chatml on an IterableDataset whose rows lack Alpaca fields ('instruction'/'output' or equivalents), or whose first row has them as None/incorrect types so the per-row conversion code throws.

Common situations: Streaming a chat-format dataset through the Alpaca->ChatML converter by mistake; Alpaca variants using 'prompt'/'response' key names instead of instruction/output; upstream schema changes after the stream was opened.

Related errors

AI-assisted analysis of unslothai/unsloth@203007d190 (2026-08-15). Data as JSON: /api/errors/45c44acdb012561d. Report an issue: GitHub.