{"record":{"id":"1d2e7667c4d355fb","repo":"sgl-project/sglang","slug":"cos-sin-shape-does-not-cover-image-tokens-and-head","errorCode":null,"errorMessage":"cos/sin shape does not cover image tokens and head_dim","messagePattern":"cos/sin shape does not cover image tokens and head_dim","errorType":"validation","errorClass":"ValueError","httpStatus":null,"severity":"error","filePath":"python/sglang/kernels/ops/diffusion/rope/hunyuan_qkv_pack_triton.py","lineNumber":183,"sourceCode":"        raise ValueError(\"QKV tensors must be CUDA bfloat16 tensors\")\n    if any(x.device != img_q.device for x in tensors):\n        raise ValueError(\"QKV tensors must be on the same CUDA device\")\n    batch, img_tokens, num_heads, head_dim = img_q.shape\n    txt_tokens = txt_q.shape[1]\n    expected_img = (batch, img_tokens, num_heads, head_dim)\n    expected_txt = (batch, txt_tokens, num_heads, head_dim)\n    if any(tuple(x.shape) != expected_img for x in (img_q, img_k, img_v)):\n        raise ValueError(\"image QKV shapes must match\")\n    if any(tuple(x.shape) != expected_txt for x in (txt_q, txt_k, txt_v)):\n        raise ValueError(\"text QKV shapes must match\")\n    if any(x.stride(-1) != 1 for x in tensors):\n        raise ValueError(\"QKV last dimensions must be contiguous\")\n    if head_dim <= 0 or head_dim > 128 or head_dim % 2:\n        raise ValueError(\"head_dim must be positive, even, and <= 128\")\n    if cos.ndim != 2 or sin.ndim != 2 or cos.shape != sin.shape:\n        raise ValueError(\"cos and sin must have matching [S, D/2] shapes\")\n    if cos.shape[0] < img_tokens or cos.shape[1] != head_dim // 2:\n        raise ValueError(\"cos/sin shape does not cover image tokens and head_dim\")\n    if not cos.is_cuda or not sin.is_cuda or cos.stride(-1) != 1 or sin.stride(-1) != 1:\n        raise ValueError(\"cos and sin must be CUDA and last-dim contiguous\")\n    if cos.device != img_q.device or sin.device != img_q.device:\n        raise ValueError(\"QKV and cos/sin tensors must be on the same CUDA device\")\n\n    total_tokens = img_tokens + txt_tokens\n    storage = torch.empty(\n        (3, batch, total_tokens, num_heads, head_dim),\n        device=img_q.device,\n        dtype=img_q.dtype,\n    )\n    args = []\n    for x in tensors:\n        args.extend((x.stride(0), x.stride(1), x.stride(2)))\n    with torch.cuda.device(img_q.device):\n        _hunyuan_qkv_rope_pack_kernel[\n            lambda meta: (\n                batch * total_tokens,","sourceCodeStart":165,"sourceCodeEnd":201,"githubUrl":"https://github.com/sgl-project/sglang/blob/0132848349585cfe6aae51c4941cbae872505f8a/python/sglang/kernels/ops/diffusion/rope/hunyuan_qkv_pack_triton.py#L165-L201","documentation":"The RoPE table must cover at least the number of image tokens (cos.shape[0] >= img_tokens) and have exactly head_dim//2 columns. Otherwise the kernel would read out of bounds when applying rotary embeddings to image tokens.","triggerScenarios":"Passing a cos/sin table sized for text-only sequence lengths, or with columns for full head_dim rather than head_dim/2, while img tokens exceed the table's rows.","commonSituations":"Mixing text RoPE tables with image-token inputs in multimodal Hunyuan pipelines, or off-by-one max-position configurations.","solutions":["Precompute cos/sin with rows >= max image token count.","Use head_dim//2 columns (half-split layout).","Check that position ids for image tokens stay within the table."],"exampleFix":"// before\ncos = torch.randn(txt_tokens, head_dim)  # wrong rows AND columns\n// after\ncos = torch.randn(max(img_tokens, txt_tokens), head_dim // 2, device='cuda')\nsin = torch.randn_like(cos)","handlingStrategy":"validation","validationCode":"assert cos.shape[0] >= img_tokens and cos.shape[1] == head_dim // 2","typeGuard":null,"tryCatchPattern":null,"preventionTips":["Size tables to the true max position including image tokens.","Use half-dim columns for split RoPE."],"tags":["rope","cos-sin","bounds-check","multimodal"],"backgroundTag":"rope-table-coverage-mismatch","analyzedSha":"0132848349585cfe6aae51c4941cbae872505f8a","analyzedAt":"2026-08-28T05:10:05.995Z","schemaVersion":2},"datasetVersion":"2026-08-28T06:17:29.519Z"}