sglang launch error when load the convert vocab

[village-way]

start script

python3 -m sglang.launch_server \
  --model /workspace/alg/llm_models/models--Qwen--Qwen2.5-Coder-32B-Instruct \
  --speculative-algorithm EAGLE \
  --speculative-draft-model-path /workspace/alg/llm_models/EAGLE-models--Qwen--Qwen2.5-coder-32B-Instruct-6635-sglang \
  --speculative-num-steps 5 \
  --speculative-eagle-topk 8 \
  --speculative-num-draft-tokens 64 \
  --mem-fraction 0.6 \
  --tensor-parallel-size 4 \
  --enable-torch-compile \
  --cuda-graph-max-bs 2 \
  --speculative-token-map /workspace/alg/FR-Spec/fr/lmh_index/Qwen2.5-Coder-32B-Instruct/freq_32768.pt

error log

[2025-03-12 11:47:50 TP0] Memory pool end. avail mem=13.09 GB
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [32,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [33,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [34,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [35,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [36,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [37,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [38,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [39,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [40,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [41,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [42,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [43,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [44,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [45,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [46,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [47,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [48,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [49,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [50,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [51,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [52,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [53,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [54,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [55,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [56,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [57,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [58,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [59,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [60,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [61,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [62,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [63,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [0,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [1,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [2,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [3,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [4,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [5,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [6,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [7,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [8,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [9,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [10,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [11,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [12,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [13,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [14,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [15,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [16,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [17,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [18,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [19,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [20,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [21,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [22,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [23,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [24,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [25,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [26,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [27,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [28,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [29,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [30,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [31,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [64,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [65,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [66,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [67,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [68,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [69,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [70,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [71,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [72,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [73,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [74,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [75,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [76,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [77,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [78,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [79,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [80,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [81,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [82,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [83,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [84,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [85,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [86,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [87,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [88,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [89,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [90,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [91,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [92,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [93,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [94,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [95,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [96,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [97,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [98,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [99,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [100,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [101,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [102,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [103,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [104,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [105,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [106,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [107,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [108,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [109,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [110,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [111,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [112,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [113,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [114,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [115,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [116,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [117,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [118,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [119,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [120,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [121,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [122,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [123,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [124,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [125,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [126,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
../aten/src/ATen/native/cuda/IndexKernel.cu:93: operator(): block: [27577,0,0], thread: [127,0,0] Assertion `-sizes[i] <= index && index < sizes[i] && "index out of bounds"` failed.
[2025-03-12 11:47:50 TP0] Scheduler hit an exception: Traceback (most recent call last):
  File "/root/anaconda3/envs/rlhf/lib/python3.10/site-packages/sglang/srt/managers/scheduler.py", line 2255, in run_scheduler_process
    scheduler = Scheduler(server_args, port_args, gpu_id, tp_rank, dp_rank)
  File "/root/anaconda3/envs/rlhf/lib/python3.10/site-packages/sglang/srt/managers/scheduler.py", line 228, in __init__
    self.draft_worker = EAGLEWorker(
  File "/root/anaconda3/envs/rlhf/lib/python3.10/site-packages/sglang/srt/speculative/eagle_worker.py", line 95, in __init__
    self.draft_model_runner.model.set_embed_and_head(embed, head)
  File "/root/anaconda3/envs/rlhf/lib/python3.10/site-packages/sglang/srt/models/qwen2.py", line 432, in set_embed_and_head
    torch.cuda.empty_cache()
  File "/root/anaconda3/envs/rlhf/lib/python3.10/site-packages/torch/cuda/memory.py", line 192, in empty_cache
    torch._C._cuda_emptyCache()
RuntimeError: CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

no fr-spec vocab with no error

convert script

python fr.py --model_name Qwen2.5-Coder-32B-Instruct \
  --model_path /workspace/alg/llm_models/models--Qwen--Qwen2.5-Coder-32B-Instruct \
  --num_lines 1000000 \
  --dataset_path /workspace/alg/llm_datasets/SlimPajama-627B

[Achazwl]

Could you share the freq_32768.pt you get to help me figure out the error?
You can send it to my email acha131441373@gmail.com

[Achazwl]

Maybe related to --tensor-parallel-size 4, where the lm_head's vocab dimension is partitioned to 4 different GPUs.

[village-way]

Could you share the freq_32768.pt you get to help me figure out the error? You can send it to my email acha131441373@gmail.com

I have send the email, you can check it.

[village-way]

Maybe related to --tensor-parallel-size 4, where the lm_head's vocab dimension is partitioned to 4 different GPUs.

I only have A100 GPU with 40GB of VRAM. Does this mean I can only set the tensor parallel size to 1? Can this vocabulary file be used for a 7B model? If so, I can try the 7B model.

[Achazwl]

Maybe related to --tensor-parallel-size 4, where the lm_head's vocab dimension is partitioned to 4 different GPUs.

I only have A100 GPU with 40GB of VRAM. Does this mean I can only set the tensor parallel size to 1? Can this vocabulary file be used for a 7B model? If so, I can try the 7B model.

Yes, if they share the same tokenizer

[Achazwl]

Does this mean I can only set the tensor parallel size to 1?

Currently, you can only use tensor parallel to 1. I'll let you know once I've fixed the bug.

[Ewan-Ding]

@Achazwl Hello, I encuntered this problem again when using the new version of sglang. Could you help me figure out what is causing it? Thank you!
The issue: sgl-project/sglang#6502

[Achazwl]

@Ewan-Ding The issue with TP>1 has been reported to sglang, but has not been addressed yet. On the other hand, the scenario where TP>1 indicates that the model size is very large (e.g., Deepseek-V3 in your case), and in such cases, the speed improvement from FR-Spec may not be very significant.

[pengchengneo]

hi @Achazwl Could you explain whats means FR-Spec ? and why large models have no significant imporvement from FR-Spec

[Achazwl]

FR-Spec is the current framework. More technical details can be seen at https://arxiv.org/abs/2502.14856.

Our paper has verified that for <=8B size models, the time bottleneck is the lm_head projection time of EAGLE-2, since EAGLE-2 would draft for multiple iterations (6 iters for typical settings) before each verification, and each draft iteration has one lm_head projection. For a typical 8B size model, such as Qwen3-8B and Llama3-8B, their lm_head has a vocabulary size at least 30x larger than the hidden size of the model itself. So 6 iterations of such a large lm_head projection would become a huge bottelneck.

FR-Spec conduct vocabulary reduction for the drafting process, generating drafts in a reduced vocabulary space to address this.

However, for larger models, while the vocabulary size remains unchanged, the width and depth of the model itself increase, which leads to a decrease in the proportion of the lmhead. As a result, the speedup ratio of our method will be reduced. (We have not yet empirically verified this hypothesis.)

[Ewan-Ding]

@Ewan-Ding The issue with TP>1 has been reported to sglang, but has not been addressed yet. On the other hand, the scenario where TP>1 indicates that the model size is very large (e.g., Deepseek-V3 in your case), and in such cases, the speed improvement from FR-Spec may not be very significant.

Thank you, your explanation is very clear.