terminate called after throwing an instance of 'std::bad_alloc' what(): std::bad_alloc #24
Description
System Info / 系統信息
========================================
System Information Diagnostic Script
Python Version: 3.10.12 (main, Aug 15 2025, 14:32:43) [GCC 11.4.0]
OS: Linux 6.6.87.2-microsoft-standard-WSL2 (#1 SMP PREEMPT_DYNAMIC Thu Jun 5 18:30:46 UTC 2025)
Platform: Linux-6.6.87.2-microsoft-standard-WSL2-x86_64-with-glibc2.35
CPU: x86_64
Physical Cores: 24
Logical Cores: 24
Total RAM: 62.53 GB
Available RAM: 28.10 GB
Transformers Version: 4.51.3
PyTorch Version: 2.9.1+cu128
CUDA Available: True
PyTorch CUDA Version: 12.8
CUDA Device Count: 1
Device 0: NVIDIA GeForce RTX 3090
Total Memory: 24.00 GB
Major/Minor: 8.6
NVCC Version (System):
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2025 NVIDIA Corporation
Built on Wed_Apr__9_19:24:57_PDT_2025
Cuda compilation tools, release 12.9, V12.9.41
Build cuda_12.9.r12.9/compiler.35813241_0
nvidia-smi Output:
Tue Dec 23 13:37:25 2025
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.82.10 Driver Version: 581.29 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 3090 On | 00000000:02:00.0 On | N/A |
| 0% 32C P8 39W / 460W | 2178MiB / 24576MiB | 2% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| 0 N/A N/A 1 C /python3.10 N/A |
| 0 N/A N/A 1 C /python3.10 N/A |
| 0 N/A N/A 1 C /python3.10 N/A |
| 0 N/A N/A 36 C /python3.10 N/A |
+-----------------------------------------------------------------------------------------+
========================================
Who can help? / 谁可以帮助到您？
No response
Information / 问题信息
- The official example scripts / 官方的示例脚本
- My own modified scripts / 我自己修改的脚本和任务
Reproduction / 复现过程
按照 requirements.txt 装好环境后，下载模型，运行 python inference.py --checkpoint_dir GLM-ASR-Nano-2512 --audio examples/example_en.wav --device cuda
就会报错，modelscope和huggingface的模型都下载测试一样的结果
Expected behavior / 期待表现
root@868442d56a08:/app# python inference.py --checkpoint_dir GLM-ASR-Nano-2512 --audio examples/example_en.wav --device cuda
[inference] start transcribe checkpoint_dir=GLM-ASR-Nano-2512 audio=examples/example_en.wav device=cuda max_new_tokens=128
[inference] cuda mem free=24.44GB total=25.77GB
[inference] load tokenizer from GLM-ASR-Nano-2512
[inference] load config from GLM-ASR-Nano-2512
You are using a model of type glmasr to instantiate a model of type Glmasr. This is not supported for all configurations of models and can yield errors.
[inference] load model with dtype=torch.float16
[inference] cuda mem free=24.44GB total=25.77GB
terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc