sgl-project/sglang · error · ValueError

speculative_adaptive_config must contain at least one intege

Error message

speculative_adaptive_config must contain at least one integer-string BS key, e.g. {"1": {"candidate_steps": [1,3,7]}}. Got keys: {list(cfg.keys())}

What it means

Validating speculative_adaptive_config: after scanning keys, no entry had an integer-string batch-size key. The config must be a dict keyed by BS as strings that parse to ints, e.g. {"1": {...}}.

Source

Thrown at python/sglang/srt/speculative/adaptive_spec_params.py:121

    bs_entries: dict[int, dict] = {}
    for key, entry in cfg.items():
        if not key.isdigit():
            continue

        steps = entry.get("candidate_steps")
        if (
            not isinstance(steps, list)
            or not steps
            or not all(isinstance(s, int) and s >= 0 for s in steps)
        ):
            raise ValueError(
                f"BS {key}: candidate_steps must be a list of non-negative ints, "
                f"got {steps!r}"
            )
        bs_entries[int(key)] = entry

    if not bs_entries:
        raise ValueError(
            "speculative_adaptive_config must contain at least one integer-string "
            'BS key, e.g. {"1": {"candidate_steps": [1,3,7]}}. '
            f"Got keys: {list(cfg.keys())}"
        )
    return cfg, bs_entries


def resolve_candidate_steps_from_config(
    cfg_path: str | None = None,
) -> list[int]:
    """Union of every BS slot's candidate steps; sizes the runtime buffers."""
    _, bs_entries = _load_adaptive_config(cfg_path)
    all_steps: set[int] = set()
    for entry in bs_entries.values():
        all_steps.update(entry["candidate_steps"])
    return sorted(all_steps)

View on GitHub (pinned to 0132848349)

Solutions

1. Rewrite the config with integer-string keys: {"1": {"candidate_steps": [1,3,7]}, "8": {...}}
2. If loading via YAML/JSON where keys become ints, serialize to a dict with str(k) keys first

Example fix

# before
speculative_adaptive_config = {8: {"candidate_steps": [1, 3]}}
# after
speculative_adaptive_config = {"8": {"candidate_steps": [1, 3]}}

assert isinstance(cfg, dict) and cfg and all(k.isdigit() for k in cfg), f'keys must be integer strings, got {list(cfg)}'

def is_valid_bs_keys(cfg) -> bool:
    return isinstance(cfg, dict) and len(cfg) > 0 and all(isinstance(k, str) and k.isdigit() for k in cfg.keys())

When it happens

Trigger: Passing an empty dict, or a dict whose keys are not integer strings (e.g. {"default": ...}, {"bs_8": ...}, or integer keys already parsed as ints by JSON/YAML loading).

Common situations: Using non-string keys in a YAML config that loads as ints; naming keys by label instead of batch size; passing the wrong structure entirely.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• BS {key}: candidate_steps must be a list of non-negative int
• DFLASH mask_token must be a non-empty string, got {mask_toke
• DSpark speculative_num_draft_tokens must be >= 2 (= gamma +
• Invalid fused KV rotary/head dim pair: rotary_dim={rotary_di
• num_kv_heads mismatch across layers for fused KV path: expec

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/fcbcf5eaec32af76. Report an issue: GitHub.