sgl-project/sglang · error · ValueError

`A_log` must be a 1D tensor.

Error message

`A_log` must be a 1D tensor.

What it means

A_log (log decay magnitudes) must be a 1D tensor of length HV for the replaySSM decode kernel — one scalar per value head, shared across tokens. Multi-dim or scalar-wrapped tensors are rejected.

Source

Thrown at python/sglang/kernels/ops/attention/fla/fused_recurrent_linear_replayssm.py:484

    Same call surface as the packed decode plus the three ring caches
    (``d_cache`` / ``k_cache`` / ``g_cache``) and the per-decode-row
    ``write_pos`` cursor.  ``initial_state`` is both the checkpoint read (h0)
    and the (flush-only) checkpoint write (ht), in place.

    Allocates nothing persistent: the caller owns the ring tensors and is
    responsible for advancing / resetting ``write_pos`` (e.g. ``(write_pos+1) %
    L`` after each step).  This is a STANDALONE kernel; the memory-pool / cache
    integration is a later phase.
    """
    if mixed_qkv.ndim != 2:
        raise ValueError(f"`mixed_qkv` must be 2D (got ndim={mixed_qkv.ndim}).")
    if mixed_qkv.stride(-1) != 1:
        raise ValueError("`mixed_qkv` must be contiguous in the last dim.")
    if b.ndim != 2:
        raise ValueError(f"`b` must be 2D (got b.ndim={b.ndim}).")
    if A_log.ndim != 1:
        raise ValueError("`A_log` must be a 1D tensor.")
    if initial_state.ndim != 4:
        raise ValueError(f"`initial_state` must be 4D (got ndim={initial_state.ndim}).")
    if not out.is_contiguous():
        raise ValueError("`out` must be contiguous.")
    if write_pos.ndim != 1 or write_pos.dtype != torch.int32:
        raise ValueError("`write_pos` must be a 1D int32 tensor.")
    if force_flush is not None and (
        force_flush.ndim != 1 or force_flush.dtype != torch.int32
    ):
        raise ValueError("`force_flush` must be a 1D int32 tensor or None.")

    B = mixed_qkv.shape[0]
    num_state_slots, HV, V, K = initial_state.shape
    qkv_dim = mixed_qkv.shape[1]
    q_dim = (qkv_dim - HV * V) // 2
    if q_dim <= 0 or q_dim % K != 0:
        raise ValueError(
            f"Invalid packed `mixed_qkv` last dim={qkv_dim} for HV={HV}, V={V}, K={K}."

View on GitHub (pinned to 0132848349)

Solutions

1. Squeeze/reshape A_log to exactly 1D: A_log.reshape(-1) (verify numel == HV)
2. Load the checkpoint parameter and squeeze() any singleton dims

Example fix

// before
A_log = ckpt['A_log']  # [1, HV]
// after
A_log = ckpt['A_log'].squeeze()  # [HV], ndim==1

A_log = A_log.reshape(-1)
assert A_log.ndim == 1 and A_log.numel() == HV

def is_1d_a_log(t: torch.Tensor, HV: int) -> bool:
    return t.ndim == 1 and t.numel() == HV

When it happens

Trigger: Passing A_log as [T, HV], [1, HV], or [B, HV] (kept a batch dim), or the raw parameter with extra dims from the checkpoint.

Common situations: Checkpoint params stored as [1, HV]; test fixtures generating batched decay tensors; reusing b's layout for A_log.

Related errors

• `mixed_qkv` must be 2D (got ndim={mixed_qkv.ndim}).
• `b` must be 2D (got b.ndim={b.ndim}).
• `initial_state` must be 4D (got ndim={initial_state.ndim}).
• `dt_bias` must have {HV * K} elements (got {dt_bias.numel()}
• Invalid packed Q size {q_dim}: must be divisible by K={K}. K

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/1b537fa8fae05859. Report an issue: GitHub.