[Bug]: streaming quantization cause higher peak memory used during model loading and post process #31805
Description
Your current environment
The output of python collect_env.py
Collecting environment information...
==============================
System Info
==============================
OS : Ubuntu 22.04.5 LTS (x86_64)
GCC version : (Ubuntu 11.4.0-1ubuntu1~22.04.2) 11.4.0
Clang version : Could not collect
CMake version : Could not collect
Libc version : glibc-2.35
==============================
PyTorch Info
==============================
PyTorch version : 2.9.0+cu129
Is debug build : False
CUDA used to build PyTorch : 12.9
ROCM used to build PyTorch : N/A
==============================
Python Environment
==============================
Python version : 3.12.12 (main, Oct 10 2025, 08:52:57) [GCC 11.4.0] (64-bit runtime)
Python platform : Linux-5.4.292-1.el8.elrepo.x86_64-x86_64-with-glibc2.35
==============================
CUDA / GPU Info
==============================
Is CUDA available : True
CUDA runtime version : 12.9.86
CUDA_MODULE_LOADING set to :
GPU models and configuration :
GPU 0: NVIDIA A100 80GB PCIe
GPU 1: NVIDIA A100 80GB PCIe
GPU 2: NVIDIA A100 80GB PCIe
GPU 3: NVIDIA A100 80GB PCIe
GPU 4: NVIDIA A100 80GB PCIe
GPU 5: NVIDIA A100 80GB PCIe
GPU 6: NVIDIA A100 80GB PCIe
Nvidia driver version : 570.133.20
cuDNN version : Could not collect
HIP runtime version : N/A
MIOpen runtime version : N/A
Is XNNPACK available : True
==============================
CPU Info
==============================
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Address sizes: 46 bits physical, 57 bits virtual
Byte Order: Little Endian
CPU(s): 128
On-line CPU(s) list: 0-127
Vendor ID: GenuineIntel
BIOS Vendor ID: Intel(R) Corporation
Model name: Intel(R) Xeon(R) Platinum 8358 CPU @ 2.60GHz
BIOS Model name: Intel(R) Xeon(R) Platinum 8358 CPU @ 2.60GHz
CPU family: 6
Model: 106
Thread(s) per core: 2
Core(s) per socket: 32
Socket(s): 2
Stepping: 6
Frequency boost: enabled
CPU max MHz: 3400.0000
CPU min MHz: 800.0000
BogoMIPS: 5200.00
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 invpcid_single ssbd mba ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local wbnoinvd dtherm ida arat pln pts avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg tme avx512_vpopcntdq la57 rdpid md_clear pconfig flush_l1d arch_capabilities
Virtualization: VT-x
L1d cache: 3 MiB (64 instances)
L1i cache: 2 MiB (64 instances)
L2 cache: 80 MiB (64 instances)
L3 cache: 96 MiB (2 instances)
NUMA node(s): 2
NUMA node0 CPU(s): 0-31,64-95
NUMA node1 CPU(s): 32-63,96-127
Vulnerability Gather data sampling: Vulnerable: No microcode
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Not affected
Vulnerability Mds: Not affected
Vulnerability Meltdown: Not affected
Vulnerability Mmio stale data: Vulnerable: Clear CPU buffers attempted, no microcode; SMT vulnerable
Vulnerability Retbleed: Not affected
Vulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2: Mitigation; Enhanced / Automatic IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected
==============================
Versions of relevant libraries
==============================
[pip3] flashinfer-python==0.5.3
[pip3] numpy==2.2.6
[pip3] nvidia-cublas-cu12==12.9.1.4
[pip3] nvidia-cuda-cupti-cu12==12.9.79
[pip3] nvidia-cuda-nvrtc-cu12==12.9.86
[pip3] nvidia-cuda-runtime-cu12==12.9.79
[pip3] nvidia-cudnn-cu12==9.10.2.21
[pip3] nvidia-cudnn-frontend==1.16.0
[pip3] nvidia-cufft-cu12==11.4.1.4
[pip3] nvidia-cufile-cu12==1.14.1.1
[pip3] nvidia-curand-cu12==10.3.10.19
[pip3] nvidia-cusolver-cu12==11.7.5.82
[pip3] nvidia-cusparse-cu12==12.5.10.65
[pip3] nvidia-cusparselt-cu12==0.7.1
[pip3] nvidia-cutlass-dsl==4.3.3
[pip3] nvidia-ml-py==13.590.44
[pip3] nvidia-nccl-cu12==2.27.5
[pip3] nvidia-nvjitlink-cu12==12.9.86
[pip3] nvidia-nvshmem-cu12==3.3.20
[pip3] nvidia-nvtx-cu12==12.9.79
[pip3] pyzmq==27.1.0
[pip3] torch==2.9.0+cu129
[pip3] torchaudio==2.9.0+cu129
[pip3] torchvision==0.24.0+cu129
[pip3] transformers==4.57.3
[pip3] triton==3.5.0
[conda] Could not collect
==============================
vLLM Info
==============================
ROCM Version : Could not collect
vLLM Version : 0.9.2rc2.dev4818+gc881db364 (git sha: c881db364)
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled
GPU Topology:
GPU0 GPU1 GPU2 GPU3 GPU4 GPU5 GPU6 CPU Affinity NUMA Affinity GPU NUMA ID
GPU0 X PIX PIX PIX SYS SYS SYS 0-31,64-95 0 N/A
GPU1 PIX X PIX PIX SYS SYS SYS 0-31,64-95 0 N/A
GPU2 PIX PIX X PIX SYS SYS SYS 0-31,64-95 0 N/A
GPU3 PIX PIX PIX X SYS SYS SYS 0-31,64-95 0 N/A
GPU4 SYS SYS SYS SYS X PIX PIX 32-63,96-127 1 N/A
GPU5 SYS SYS SYS SYS PIX X PIX 32-63,96-127 1 N/A
GPU6 SYS SYS SYS SYS PIX PIX X 32-63,96-127 1 N/A
Legend:
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
==============================
Environment Variables
==============================
NVIDIA_VISIBLE_DEVICES=all
NVIDIA_REQUIRE_CUDA=cuda>=12.9 brand=unknown,driver>=535,driver<536 brand=grid,driver>=535,driver<536 brand=tesla,driver>=535,driver<536 brand=nvidia,driver>=535,driver<536 brand=quadro,driver>=535,driver<536 brand=quadrortx,driver>=535,driver<536 brand=nvidiartx,driver>=535,driver<536 brand=vapps,driver>=535,driver<536 brand=vpc,driver>=535,driver<536 brand=vcs,driver>=535,driver<536 brand=vws,driver>=535,driver<536 brand=cloudgaming,driver>=535,driver<536 brand=unknown,driver>=550,driver<551 brand=grid,driver>=550,driver<551 brand=tesla,driver>=550,driver<551 brand=nvidia,driver>=550,driver<551 brand=quadro,driver>=550,driver<551 brand=quadrortx,driver>=550,driver<551 brand=nvidiartx,driver>=550,driver<551 brand=vapps,driver>=550,driver<551 brand=vpc,driver>=550,driver<551 brand=vcs,driver>=550,driver<551 brand=vws,driver>=550,driver<551 brand=cloudgaming,driver>=550,driver<551 brand=unknown,driver>=560,driver<561 brand=grid,driver>=560,driver<561 brand=tesla,driver>=560,driver<561 brand=nvidia,driver>=560,driver<561 brand=quadro,driver>=560,driver<561 brand=quadrortx,driver>=560,driver<561 brand=nvidiartx,driver>=560,driver<561 brand=vapps,driver>=560,driver<561 brand=vpc,driver>=560,driver<561 brand=vcs,driver>=560,driver<561 brand=vws,driver>=560,driver<561 brand=cloudgaming,driver>=560,driver<561 brand=unknown,driver>=565,driver<566 brand=grid,driver>=565,driver<566 brand=tesla,driver>=565,driver<566 brand=nvidia,driver>=565,driver<566 brand=quadro,driver>=565,driver<566 brand=quadrortx,driver>=565,driver<566 brand=nvidiartx,driver>=565,driver<566 brand=vapps,driver>=565,driver<566 brand=vpc,driver>=565,driver<566 brand=vcs,driver>=565,driver<566 brand=vws,driver>=565,driver<566 brand=cloudgaming,driver>=565,driver<566 brand=unknown,driver>=570,driver<571 brand=grid,driver>=570,driver<571 brand=tesla,driver>=570,driver<571 brand=nvidia,driver>=570,driver<571 brand=quadro,driver>=570,driver<571 brand=quadrortx,driver>=570,driver<571 brand=nvidiartx,driver>=570,driver<571 brand=vapps,driver>=570,driver<571 brand=vpc,driver>=570,driver<571 brand=vcs,driver>=570,driver<571 brand=vws,driver>=570,driver<571 brand=cloudgaming,driver>=570,driver<571
NVIDIA_DRIVER_CAPABILITIES=compute,utility
VLLM_USAGE_SOURCE=production-docker-image
CUDA_VERSION=12.9.1
LD_LIBRARY_PATH=/usr/local/nvidia/lib64:/usr/local/cuda/lib64:/usr/local/cuda/lib64
PYTORCH_NVML_BASED_CUDA_CHECK=1
TORCHINDUCTOR_COMPILE_THREADS=1
🐛 Describe the bug
I simply print the memory allocation status during model loading w/ and w/o patched_weight_loader and can see the peak memory used becomes higher and eventually failed to load.
Device: A100
CMD:
python3 examples/offline_inference/basic/generate.py --model Qwen/Qwen3-30B-A3B --enforce-eager --dtype=bfloat16 --block-size=64 --max_model_len=2048 --gpu-memory-utilization=0.8 --trust-remote-code --quantization=fp8
w/ patched_weight_loader:
(EngineCore_DP0 pid=5765) Memory allocated: 1.7814621925354004
(EngineCore_DP0 pid=5765) Memory allocated: 2.9421215057373047
(EngineCore_DP0 pid=5765) Memory allocated: 4.102780818939209
(EngineCore_DP0 pid=5765) Memory allocated: 5.263440132141113
(EngineCore_DP0 pid=5765) Memory allocated: 6.424099445343018
(EngineCore_DP0 pid=5765) Memory allocated: 7.584758758544922
(EngineCore_DP0 pid=5765) Memory allocated: 8.745418071746826
(EngineCore_DP0 pid=5765) Memory allocated: 9.90607738494873
(EngineCore_DP0 pid=5765) Memory allocated: 11.066736698150635
(EngineCore_DP0 pid=5765) Memory allocated: 12.227396011352539
(EngineCore_DP0 pid=5765) Memory allocated: 13.388055324554443
(EngineCore_DP0 pid=5765) Memory allocated: 14.548714637756348
(EngineCore_DP0 pid=5765) Memory allocated: 15.709373950958252
(EngineCore_DP0 pid=5765) Memory allocated: 16.870033264160156
(EngineCore_DP0 pid=5765) Memory allocated: 18.03069257736206
(EngineCore_DP0 pid=5765) Memory allocated: 19.191351890563965
(EngineCore_DP0 pid=5765) Memory allocated: 20.35201120376587
(EngineCore_DP0 pid=5765) Memory allocated: 21.512670516967773
(EngineCore_DP0 pid=5765) Memory allocated: 22.673329830169678
(EngineCore_DP0 pid=5765) Memory allocated: 23.833989143371582
(EngineCore_DP0 pid=5765) Memory allocated: 24.994648456573486
(EngineCore_DP0 pid=5765) Memory allocated: 26.15530776977539
(EngineCore_DP0 pid=5765) Memory allocated: 27.315967082977295
(EngineCore_DP0 pid=5765) Memory allocated: 28.4766263961792
(EngineCore_DP0 pid=5765) Memory allocated: 29.637285709381104
(EngineCore_DP0 pid=5765) Memory allocated: 30.797945022583008
(EngineCore_DP0 pid=5765) Memory allocated: 31.958604335784912
(EngineCore_DP0 pid=5765) Memory allocated: 33.119263648986816
(EngineCore_DP0 pid=5765) Memory allocated: 34.27992296218872
(EngineCore_DP0 pid=5765) Memory allocated: 35.440582275390625
(EngineCore_DP0 pid=5765) Memory allocated: 36.60124158859253
(EngineCore_DP0 pid=5765) Memory allocated: 37.761900901794434
(EngineCore_DP0 pid=5765) Memory allocated: 38.92256021499634
(EngineCore_DP0 pid=5765) Memory allocated: 40.08321952819824
(EngineCore_DP0 pid=5765) Memory allocated: 41.24387884140015
(EngineCore_DP0 pid=5765) Memory allocated: 42.40453815460205
(EngineCore_DP0 pid=5765) Memory allocated: 43.565197467803955
(EngineCore_DP0 pid=5765) Memory allocated: 44.72585678100586
(EngineCore_DP0 pid=5765) Memory allocated: 45.886516094207764
(EngineCore_DP0 pid=5765) Memory allocated: 47.04717540740967
(EngineCore_DP0 pid=5765) Memory allocated: 48.20783472061157
(EngineCore_DP0 pid=5765) Memory allocated: 49.36849403381348
(EngineCore_DP0 pid=5765) Memory allocated: 50.52915334701538
(EngineCore_DP0 pid=5765) Memory allocated: 51.689812660217285
(EngineCore_DP0 pid=5765) Memory allocated: 52.85047197341919
(EngineCore_DP0 pid=5765) Memory allocated: 54.011131286621094
(EngineCore_DP0 pid=5765) Memory allocated: 55.171790599823
(EngineCore_DP0 pid=5765) Memory allocated: 56.3324499130249
(EngineCore_DP0 pid=5765) INFO 01-05 22:46:26 [weight_utils.py:510] Time spent downloading weights for Qwen/Qwen3-30B-A3B: 0.576616 seconds
Loading safetensors checkpoint shards: 0% Completed | 0/16 [00:00<?, ?it/s]
Loading safetensors checkpoint shards: 6% Completed | 1/16 [00:00<00:02, 5.20it/s]
(EngineCore_DP0 pid=5765) WARNING 01-05 22:46:27 [marlin_utils_fp8.py:214] Your GPU does not have native support for FP8 computation but FP8 quantization is being used. Weight-only FP8 compression will be used leveraging the Marlin kernel. This may degrade performance for compute-heavy workloads.
(EngineCore_DP0 pid=5765) processing... memory allocated: 58.03888416290283
(EngineCore_DP0 pid=5765) processing... memory allocated: 58.62007808685303
Loading safetensors checkpoint shards: 12% Completed | 2/16 [00:01<00:09, 1.43it/s]
(EngineCore_DP0 pid=5765) processing... memory allocated: 59.219109535217285
(EngineCore_DP0 pid=5765) processing... memory allocated: 59.80030345916748
Loading safetensors checkpoint shards: 19% Completed | 3/16 [00:02<00:11, 1.17it/s]
(EngineCore_DP0 pid=5765) processing... memory allocated: 60.39933490753174
(EngineCore_DP0 pid=5765) processing... memory allocated: 60.980528831481934
(EngineCore_DP0 pid=5765) processing... memory allocated: 61.56172275543213
Loading safetensors checkpoint shards: 25% Completed | 4/16 [00:03<00:12, 1.02s/it]
(EngineCore_DP0 pid=5765) processing... memory allocated: 62.16075420379639
(EngineCore_DP0 pid=5765) processing... memory allocated: 62.74194812774658
(EngineCore_DP0 pid=5765) processing... memory allocated: 63.32314205169678
Loading safetensors checkpoint shards: 31% Completed | 5/16 [00:04<00:11, 1.09s/it]
(EngineCore_DP0 pid=5765) processing... memory allocated: 63.90433597564697
(EngineCore_DP0 pid=5765) processing... memory allocated: 64.48552989959717
Loading safetensors checkpoint shards: 38% Completed | 6/16 [00:05<00:11, 1.12s/it]
(EngineCore_DP0 pid=5765) processing... memory allocated: 65.08456134796143
(EngineCore_DP0 pid=5765) processing... memory allocated: 65.64791774749756
(EngineCore_DP0 pid=5765) processing... memory allocated: 66.22911167144775
(EngineCore_DP0 pid=5765) processing... memory allocated: 66.81030559539795
Loading safetensors checkpoint shards: 44% Completed | 7/16 [00:07<00:10, 1.15s/it]
(EngineCore_DP0 pid=5765) processing... memory allocated: 67.39149951934814
(EngineCore_DP0 pid=5765) processing... memory allocated: 67.95485591888428
(EngineCore_DP0 pid=5765) processing... memory allocated: 68.53604984283447
Loading safetensors checkpoint shards: 50% Completed | 8/16 [00:08<00:09, 1.20s/it]
(EngineCore_DP0 pid=5765) processing... memory allocated: 69.13508129119873
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [gpu_model_runner.py:3797] Failed to load model - not enough GPU memory. Try lowering --gpu-memory-utilization to free memory for weights, increasing --tensor-parallel-size, or using --quantization. See https://docs.vllm.ai/en/latest/configuration/conserving_memory/ for more tips. (original error: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 79.25 GiB of which 157.38 MiB is free. Process 1952521 has 8.96 GiB memory in use. Process 1956684 has 70.13 GiB memory in use. Of the allocated memory 69.53 GiB is allocated by PyTorch, and 100.41 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables))
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] EngineCore failed to start.
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] Traceback (most recent call last):
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/engine/core.py", line 868, in run_engine_core
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] engine_core = EngineCoreProc(*args, **kwargs)
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/engine/core.py", line 648, in __init__
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] super().__init__(
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/engine/core.py", line 104, in __init__
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self.model_executor = executor_class(vllm_config)
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] ^^^^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/executor/abstract.py", line 101, in __init__
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self._init_executor()
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/executor/uniproc_executor.py", line 48, in _init_executor
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self.driver_worker.load_model()
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_worker.py", line 277, in load_model
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self.model_runner.load_model(eep_scale_up=eep_scale_up)
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_model_runner.py", line 3798, in load_model
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] raise e
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_model_runner.py", line 3721, in load_model
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self.model = model_loader.load_model(
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] ^^^^^^^^^^^^^^^^^^^^^^^^
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/model_loader/base_loader.py", line 55, in load_model
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] self.load_weights(model, model_config)
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] File "/usr/local/lib/python3.12/dist-packages/vllm/model_executor/model_loader/default_loader.py", line 305, in load_weights
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] loaded_weights = model.load_weights(self.get_all_weights(model_config, model))
(EngineCore_DP0 pid=5765) ERROR 01-05 22:46:36 [core.py:877] ^^^^^^^^^^^^^^^
w/o patched_weight_loader:
(EngineCore_DP0 pid=6036) INFO 01-06 03:54:13 [layer.py:366] Enabled separate cuda stream for MoE shared_experts
(EngineCore_DP0 pid=6036) INFO 01-06 03:54:13 [fp8.py:168] Using Marlin backend for FP8 MoE
(EngineCore_DP0 pid=6036) Memory allocated: 1.7814621925354004
(EngineCore_DP0 pid=6036) Memory allocated: 2.9421215057373047
(EngineCore_DP0 pid=6036) Memory allocated: 4.102780818939209
(EngineCore_DP0 pid=6036) Memory allocated: 5.263440132141113
(EngineCore_DP0 pid=6036) Memory allocated: 6.424099445343018
(EngineCore_DP0 pid=6036) Memory allocated: 7.584758758544922
(EngineCore_DP0 pid=6036) Memory allocated: 8.745418071746826
(EngineCore_DP0 pid=6036) Memory allocated: 9.90607738494873
(EngineCore_DP0 pid=6036) Memory allocated: 11.066736698150635
(EngineCore_DP0 pid=6036) Memory allocated: 12.227396011352539
(EngineCore_DP0 pid=6036) Memory allocated: 13.388055324554443
(EngineCore_DP0 pid=6036) Memory allocated: 14.548714637756348
(EngineCore_DP0 pid=6036) Memory allocated: 15.709373950958252
(EngineCore_DP0 pid=6036) Memory allocated: 16.870033264160156
(EngineCore_DP0 pid=6036) Memory allocated: 18.03069257736206
(EngineCore_DP0 pid=6036) Memory allocated: 19.191351890563965
(EngineCore_DP0 pid=6036) Memory allocated: 20.35201120376587
(EngineCore_DP0 pid=6036) Memory allocated: 21.512670516967773
(EngineCore_DP0 pid=6036) Memory allocated: 22.673329830169678
(EngineCore_DP0 pid=6036) Memory allocated: 23.833989143371582
(EngineCore_DP0 pid=6036) Memory allocated: 24.994648456573486
(EngineCore_DP0 pid=6036) Memory allocated: 26.15530776977539
(EngineCore_DP0 pid=6036) Memory allocated: 27.315967082977295
(EngineCore_DP0 pid=6036) Memory allocated: 28.4766263961792
(EngineCore_DP0 pid=6036) Memory allocated: 29.637285709381104
(EngineCore_DP0 pid=6036) Memory allocated: 30.797945022583008
(EngineCore_DP0 pid=6036) Memory allocated: 31.958604335784912
(EngineCore_DP0 pid=6036) Memory allocated: 33.119263648986816
(EngineCore_DP0 pid=6036) Memory allocated: 34.27992296218872
(EngineCore_DP0 pid=6036) Memory allocated: 35.440582275390625
(EngineCore_DP0 pid=6036) Memory allocated: 36.60124158859253
(EngineCore_DP0 pid=6036) Memory allocated: 37.761900901794434
(EngineCore_DP0 pid=6036) Memory allocated: 38.92256021499634
(EngineCore_DP0 pid=6036) Memory allocated: 40.08321952819824
(EngineCore_DP0 pid=6036) Memory allocated: 41.24387884140015
(EngineCore_DP0 pid=6036) Memory allocated: 42.40453815460205
(EngineCore_DP0 pid=6036) Memory allocated: 43.565197467803955
(EngineCore_DP0 pid=6036) Memory allocated: 44.72585678100586
(EngineCore_DP0 pid=6036) Memory allocated: 45.886516094207764
(EngineCore_DP0 pid=6036) Memory allocated: 47.04717540740967
(EngineCore_DP0 pid=6036) Memory allocated: 48.20783472061157
(EngineCore_DP0 pid=6036) Memory allocated: 49.36849403381348
(EngineCore_DP0 pid=6036) Memory allocated: 50.52915334701538
(EngineCore_DP0 pid=6036) Memory allocated: 51.689812660217285
(EngineCore_DP0 pid=6036) Memory allocated: 52.85047197341919
(EngineCore_DP0 pid=6036) Memory allocated: 54.011131286621094
(EngineCore_DP0 pid=6036) Memory allocated: 55.171790599823
(EngineCore_DP0 pid=6036) Memory allocated: 56.3324499130249
(EngineCore_DP0 pid=6036) INFO 01-06 03:54:17 [weight_utils.py:510] Time spent downloading weights for Qwen/Qwen3-30B-A3B: 0.652724 seconds
Loading safetensors checkpoint shards: 0% Completed | 0/16 [00:00<?, ?it/s]
Loading safetensors checkpoint shards: 6% Completed | 1/16 [00:00<00:02, 5.84it/s]
Loading safetensors checkpoint shards: 12% Completed | 2/16 [00:00<00:07, 1.94it/s]
Loading safetensors checkpoint shards: 19% Completed | 3/16 [00:01<00:08, 1.54it/s]
Loading safetensors checkpoint shards: 25% Completed | 4/16 [00:02<00:08, 1.37it/s]
Loading safetensors checkpoint shards: 31% Completed | 5/16 [00:03<00:08, 1.30it/s]
Loading safetensors checkpoint shards: 38% Completed | 6/16 [00:04<00:07, 1.26it/s]
Loading safetensors checkpoint shards: 44% Completed | 7/16 [00:05<00:07, 1.23it/s]
Loading safetensors checkpoint shards: 50% Completed | 8/16 [00:05<00:06, 1.21it/s]
Loading safetensors checkpoint shards: 56% Completed | 9/16 [00:06<00:05, 1.20it/s]
Loading safetensors checkpoint shards: 62% Completed | 10/16 [00:07<00:05, 1.19it/s]
Loading safetensors checkpoint shards: 69% Completed | 11/16 [00:08<00:04, 1.19it/s]
Loading safetensors checkpoint shards: 75% Completed | 12/16 [00:09<00:03, 1.19it/s]
Loading safetensors checkpoint shards: 81% Completed | 13/16 [00:10<00:02, 1.19it/s]
Loading safetensors checkpoint shards: 88% Completed | 14/16 [00:11<00:01, 1.18it/s]
Loading safetensors checkpoint shards: 94% Completed | 15/16 [00:11<00:00, 1.18it/s]
Loading safetensors checkpoint shards: 100% Completed | 16/16 [00:12<00:00, 1.20it/s]
Loading safetensors checkpoint shards: 100% Completed | 16/16 [00:12<00:00, 1.26it/s]
(EngineCore_DP0 pid=6036)
(EngineCore_DP0 pid=6036) INFO 01-06 03:54:31 [default_loader.py:308] Loading weights took 12.83 seconds
(EngineCore_DP0 pid=6036) WARNING 01-06 03:54:31 [marlin_utils_fp8.py:96] Your GPU does not have native support for FP8 computation but FP8 quantization is being used. Weight-only FP8 compression will be used leveraging the Marlin kernel. This may degrade performance for compute-heavy workloads.
(EngineCore_DP0 pid=6036) processing... memory allocated: 56.89631795883179
(EngineCore_DP0 pid=6036) processing... memory allocated: 56.31686449050903
(EngineCore_DP0 pid=6036) processing... memory allocated: 55.73741102218628
(EngineCore_DP0 pid=6036) processing... memory allocated: 55.157957553863525
(EngineCore_DP0 pid=6036) processing... memory allocated: 54.57850408554077
(EngineCore_DP0 pid=6036) processing... memory allocated: 53.99905061721802
(EngineCore_DP0 pid=6036) processing... memory allocated: 53.419597148895264
(EngineCore_DP0 pid=6036) processing... memory allocated: 52.84014368057251
(EngineCore_DP0 pid=6036) processing... memory allocated: 52.260690212249756
(EngineCore_DP0 pid=6036) processing... memory allocated: 51.681236743927
(EngineCore_DP0 pid=6036) processing... memory allocated: 51.10178327560425
(EngineCore_DP0 pid=6036) processing... memory allocated: 50.522329807281494
(EngineCore_DP0 pid=6036) processing... memory allocated: 49.94287633895874
(EngineCore_DP0 pid=6036) processing... memory allocated: 49.363422870635986
(EngineCore_DP0 pid=6036) processing... memory allocated: 48.78396940231323
(EngineCore_DP0 pid=6036) processing... memory allocated: 48.20451593399048
(EngineCore_DP0 pid=6036) processing... memory allocated: 47.625062465667725
(EngineCore_DP0 pid=6036) processing... memory allocated: 47.04560899734497
(EngineCore_DP0 pid=6036) processing... memory allocated: 46.46615552902222
(EngineCore_DP0 pid=6036) processing... memory allocated: 45.88670206069946
(EngineCore_DP0 pid=6036) processing... memory allocated: 45.30724859237671
(EngineCore_DP0 pid=6036) processing... memory allocated: 44.727795124053955
(EngineCore_DP0 pid=6036) processing... memory allocated: 44.1483416557312
(EngineCore_DP0 pid=6036) processing... memory allocated: 43.56888818740845
(EngineCore_DP0 pid=6036) processing... memory allocated: 42.98943471908569
(EngineCore_DP0 pid=6036) processing... memory allocated: 42.40998125076294
(EngineCore_DP0 pid=6036) processing... memory allocated: 41.830527782440186
(EngineCore_DP0 pid=6036) processing... memory allocated: 41.25107431411743
(EngineCore_DP0 pid=6036) processing... memory allocated: 40.67162084579468
(EngineCore_DP0 pid=6036) processing... memory allocated: 40.092167377471924
(EngineCore_DP0 pid=6036) processing... memory allocated: 39.51271390914917
(EngineCore_DP0 pid=6036) processing... memory allocated: 38.933260440826416
(EngineCore_DP0 pid=6036) processing... memory allocated: 38.35380697250366
(EngineCore_DP0 pid=6036) processing... memory allocated: 37.77435350418091
(EngineCore_DP0 pid=6036) processing... memory allocated: 37.194900035858154
(EngineCore_DP0 pid=6036) processing... memory allocated: 36.6154465675354
(EngineCore_DP0 pid=6036) processing... memory allocated: 36.03599309921265
(EngineCore_DP0 pid=6036) processing... memory allocated: 35.45653963088989
(EngineCore_DP0 pid=6036) processing... memory allocated: 34.87708616256714
(EngineCore_DP0 pid=6036) processing... memory allocated: 34.297632694244385
(EngineCore_DP0 pid=6036) processing... memory allocated: 33.71817922592163
(EngineCore_DP0 pid=6036) processing... memory allocated: 33.13872575759888
(EngineCore_DP0 pid=6036) processing... memory allocated: 32.55927228927612
(EngineCore_DP0 pid=6036) processing... memory allocated: 31.97981882095337
(EngineCore_DP0 pid=6036) processing... memory allocated: 31.400365352630615
(EngineCore_DP0 pid=6036) processing... memory allocated: 30.82091188430786
(EngineCore_DP0 pid=6036) processing... memory allocated: 30.241458415985107
(EngineCore_DP0 pid=6036) processing... memory allocated: 29.662004947662354
I think this feature is not working as expected.
Before submitting a new issue...
- Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.