{"record":{"id":"1b537fa8fae05859","repo":"sgl-project/sglang","slug":"a-log-must-be-a-1d-tensor","errorCode":null,"errorMessage":"`A_log` must be a 1D tensor.","messagePattern":"`A_log` must be a 1D tensor\\.","errorType":"validation","errorClass":"ValueError","httpStatus":null,"severity":"error","filePath":"python/sglang/kernels/ops/attention/fla/fused_recurrent_linear_replayssm.py","lineNumber":484,"sourceCode":"\n    Same call surface as the packed decode plus the three ring caches\n    (``d_cache`` / ``k_cache`` / ``g_cache``) and the per-decode-row\n    ``write_pos`` cursor.  ``initial_state`` is both the checkpoint read (h0)\n    and the (flush-only) checkpoint write (ht), in place.\n\n    Allocates nothing persistent: the caller owns the ring tensors and is\n    responsible for advancing / resetting ``write_pos`` (e.g. ``(write_pos+1) %\n    L`` after each step).  This is a STANDALONE kernel; the memory-pool / cache\n    integration is a later phase.\n    \"\"\"\n    if mixed_qkv.ndim != 2:\n        raise ValueError(f\"`mixed_qkv` must be 2D (got ndim={mixed_qkv.ndim}).\")\n    if mixed_qkv.stride(-1) != 1:\n        raise ValueError(\"`mixed_qkv` must be contiguous in the last dim.\")\n    if b.ndim != 2:\n        raise ValueError(f\"`b` must be 2D (got b.ndim={b.ndim}).\")\n    if A_log.ndim != 1:\n        raise ValueError(\"`A_log` must be a 1D tensor.\")\n    if initial_state.ndim != 4:\n        raise ValueError(f\"`initial_state` must be 4D (got ndim={initial_state.ndim}).\")\n    if not out.is_contiguous():\n        raise ValueError(\"`out` must be contiguous.\")\n    if write_pos.ndim != 1 or write_pos.dtype != torch.int32:\n        raise ValueError(\"`write_pos` must be a 1D int32 tensor.\")\n    if force_flush is not None and (\n        force_flush.ndim != 1 or force_flush.dtype != torch.int32\n    ):\n        raise ValueError(\"`force_flush` must be a 1D int32 tensor or None.\")\n\n    B = mixed_qkv.shape[0]\n    num_state_slots, HV, V, K = initial_state.shape\n    qkv_dim = mixed_qkv.shape[1]\n    q_dim = (qkv_dim - HV * V) // 2\n    if q_dim <= 0 or q_dim % K != 0:\n        raise ValueError(\n            f\"Invalid packed `mixed_qkv` last dim={qkv_dim} for HV={HV}, V={V}, K={K}.\"","sourceCodeStart":466,"sourceCodeEnd":502,"githubUrl":"https://github.com/sgl-project/sglang/blob/0132848349585cfe6aae51c4941cbae872505f8a/python/sglang/kernels/ops/attention/fla/fused_recurrent_linear_replayssm.py#L466-L502","documentation":"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.","triggerScenarios":"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.","commonSituations":"Checkpoint params stored as [1, HV]; test fixtures generating batched decay tensors; reusing b's layout for A_log.","solutions":["Squeeze/reshape A_log to exactly 1D: A_log.reshape(-1) (verify numel == HV)","Load the checkpoint parameter and squeeze() any singleton dims"],"exampleFix":"// before\nA_log = ckpt['A_log']  # [1, HV]\n// after\nA_log = ckpt['A_log'].squeeze()  # [HV], ndim==1","handlingStrategy":"validation","validationCode":"A_log = A_log.reshape(-1)\nassert A_log.ndim == 1 and A_log.numel() == HV","typeGuard":"def is_1d_a_log(t: torch.Tensor, HV: int) -> bool:\n    return t.ndim == 1 and t.numel() == HV","tryCatchPattern":null,"preventionTips":["Squeeze checkpoint params at load time","Add unit tests for param layouts"],"tags":["pytorch","tensor-shape","replayssm","decay"],"backgroundTag":"tensor-shape-mismatch","analyzedSha":"0132848349585cfe6aae51c4941cbae872505f8a","analyzedAt":"2026-08-28T05:10:05.995Z","schemaVersion":2},"datasetVersion":"2026-08-28T06:17:29.519Z"}