CPU training crashes on WSL2 due to CUDA_VISIBLE_DEVICES="-1" (double free detected in tcache 2)

### Search before asking - [x] I have searched https://github.com/ultralytics/yolov5/issues and did not find a similar report. ### Project area Training ### Bug **Bug description** When running YOLOv5 on **WSL2** with `--device cpu`, the program crashes immediately with a C++ memory error: `free(): double free detected in tcache 2`. It is mainly caused by the WSL2 CUDA driver (libcuda.so) when setting os.environ["CUDA_VISIBLE_DEVICES"] to an unavailable value (such as "-1" or ""). Below is a detailed report generated by Gemini to demonstrate the root cause and a minimal reproducible example. The root cause is in `utils/torch_utils.py` inside the `select_device()` function. To force CPU mode, YOLOv5 sets `os.environ["CUDA_VISIBLE_DEVICES"] = "-1"`. However, in the WSL2 environment, explicitly setting this environment variable to an invalid device string (`"-1"`, `""`, or `"abc"`) and then calling `torch.cuda.is_available()` triggers a fatal bug in the underlying WSL2 CUDA driver (`libcuda.so`), causing a double free abort. **Minimal Reproducible Example (MRE)** This bug is entirely reproducible outside of YOLOv5 with just 4 lines of Python code in a WSL2 terminal: ```python import torch import os # Setting it to "-1" (like YOLOv5 does for CPU), "", or "abc" os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # This triggers the driver initialization and immediately crashes torch.cuda.is_available() ``` **Terminal Output:** ```text free(): double free detected in tcache 2 Aborted (core dumped) ``` **Where it happens in YOLOv5** In `utils/torch_utils.py`, roughly at line 121: ```python if cpu or mps: os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # <--- THIS LINE CAUSES THE CRASH IN WSL2 ``` **Workaround / Proposed Solution** For WSL2 users, deleting or commenting out `os.environ["CUDA_VISIBLE_DEVICES"] = "-1"` perfectly resolves the crash, and the program happily runs on CPU. And here is the complete call stack when error occured. ``` train: Scanning /home/polaris/code/python/hw_yolov5/VOCData/dataSet_path/train.cache... 413 images, 0 backgrounds, 0 corrupt: 100%|██████████| 413/ [New Thread 0x7ffe8d7ff6c0 (LWP 31972)] [New Thread 0x7ffe8cbff6c0 (LWP 31973)] [Thread 0x7ffe8cbff6c0 (LWP 31973) exited] free(): double free detected in tcache 2 Thread 1 "python" received signal SIGABRT, Aborted. __pthread_kill_implementation (no_tid=0, signo=6, threadid=<optimized out>) at ./nptl/pthread_kill.c:44 warning: 44 ./nptl/pthread_kill.c: No such file or directory (gdb) bt #0 __pthread_kill_implementation (no_tid=0, signo=6, threadid=<optimized out>) at ./nptl/pthread_kill.c:44 #1 __pthread_kill_internal (signo=6, threadid=<optimized out>) at ./nptl/pthread_kill.c:78 #2 __GI___pthread_kill (threadid=<optimized out>, signo=signo@entry=6) at ./nptl/pthread_kill.c:89 #3 0x00007ffff7c4527e in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26 #4 0x00007ffff7c288ff in __GI_abort () at ./stdlib/abort.c:79 #5 0x00007ffff7c297b6 in __libc_message_impl (fmt=fmt@entry=0x7ffff7dce8d7 "%s\n") at ../sysdeps/posix/libc_fatal.c:134 #6 0x00007ffff7ca8ff5 in malloc_printerr (str=str@entry=0x7ffff7dd1bf0 "free(): double free detected in tcache 2") at ./malloc/malloc.c:5775 #7 0x00007ffff7cab55f in _int_free (av=0x7ffff7e03ac0 <main_arena>, p=<optimized out>, have_lock=0) at ./malloc/malloc.c:4541 #8 0x00007ffff7caddce in __GI___libc_free (mem=0x7b218a0) at ./malloc/malloc.c:3398 #9 0x00007ffe8dd16005 in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 --Type <RET> for more, q to quit, c to continue without paging--c #10 0x00007ffe8dd2a0a9 in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #11 0x00007ffe8db5429d in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #12 0x00007ffe8db55695 in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #13 0x00007ffe8db155b4 in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #14 0x00007ffe8dbaeec2 in ?? () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #15 0x00007ffe8db3aa6b in cuInit () from /usr/lib/wsl/drivers/nvcvui.inf_amd64_87477ececce1e1c5/libcuda.so.1.1 #16 0x00007ffff42332e1 in ?? () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/../../nvidia/cuda_runtime/lib/libcudart.so.12 #17 0x00007ffff4236318 in ?? () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/../../nvidia/cuda_runtime/lib/libcudart.so.12 #18 0x00007ffff7ca1ed3 in __pthread_once_slow (once_control=0x7ffff44a6148, init_routine=0x7ffff42362d0) at ./nptl/pthread_once.c:116 #19 0x00007ffff42786e9 in ?? () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/../../nvidia/cuda_runtime/lib/libcudart.so.12 #20 0x00007ffff4227a67 in ?? () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/../../nvidia/cuda_runtime/lib/libcudart.so.12 #21 0x00007ffff4249811 in cudaGetDeviceCount () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/../../nvidia/cuda_runtime/lib/libcudart.so.12 #22 0x00007ffff458504a in c10::cuda::device_count() () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/libc10_cuda.so #23 0x00007fffd7e02163 in THCPModule_getDeviceCount_wrap(_object*, _object*) () from /home/polaris/programs/miniconda3/envs/yolov5/lib/python3.8/site-packages/torch/lib/libtorch_python.so #24 0x00000000004e64b5 in cfunction_vectorcall_NOARGS (func=0x7fffd84a6e50, args=<optimized out>, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/methodobject.c:463 #25 0x00000000004dcdc9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d43668, callable=0x7fffd84a6e50) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #26 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #27 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3469 #28 0x00000000004e8486 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffed5d434f0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #29 function_code_fastcall (globals=<optimized out>, nargs=0, args=<optimized out>, co=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:284 #30 _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ffed5d233b8, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:411 #31 0x00000000004dcdc9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d233b8, callable=0x7ffff379f280) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #32 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #33 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3469 #34 0x00000000004f5556 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffed5d23210) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #35 function_code_fastcall (globals=<optimized out>, nargs=2, args=<optimized out>, co=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:284 #36 _PyFunction_Vectorcall (kwnames=<optimized out>, nargsf=<optimized out>, stack=0x7b003a8, func=0x7fffea3b6550) at /usr/local/src/conda/python-3.8.20/Objects/call.c:411 #37 _PyObject_Vectorcall (kwnames=<optimized out>, nargsf=<optimized out>, args=0x7b003a8, callable=0x7fffea3b6550) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #38 method_vectorcall (method=<optimized out>, args=0x7b003b0, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/classobject.c:60 #39 0x00000000004dcdc9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7b003b0, callable=0x7ffed5cda300) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #40 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #41 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3469 #42 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0x7b001e0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #43 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=locals@entry=0x0, args=args@entry=0x7fffffffc3c0, argcount=argcount@entry=2, kwnames=kwnames@entry=0x0, kwargs=0x7fffffffc3d0, kwcount=<optimized out>, kwstep=1, defs=0x0, defcount=<optimized out>, kwdefs=0x0, closure=0x7fffea3b28b0, name=0x7ffff7e1d530, qualname=0x7fffea4292d0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #44 0x00000000004e081b in _PyFunction_Vectorcall (kwnames=0x0, nargsf=<optimized out>, stack=0x7fffffffc3c0, func=0x7fffea3b6af0) at /usr/local/src/conda/python-3.8.20/Objects/call.c:436 #45 _PyObject_FastCallDict (callable=0x7fffea3b6af0, args=0x7fffffffc3c0, nargsf=<optimized out>, kwargs=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:96 #46 0x00000000004f2143 in _PyObject_Call_Prepend (kwargs=0x0, args=0x7ffed5c12340, obj=0x7ffed5d5d370, callable=0x7fffea3b6af0) at /usr/local/src/conda/python-3.8.20/Objects/call.c:888 #47 slot_tp_init (self=0x7ffed5d5d370, args=0x7ffed5c12340, kwds=0x0) at /usr/local/src/conda/python-3.8.20/Objects/typeobject.c:6790 #48 0x00000000004e10e3 in type_call (kwds=0x0, args=0x7ffed5c12340, type=0x4addfb0) at /usr/local/src/conda/python-3.8.20/Objects/typeobject.c:994 #49 _PyObject_MakeTpCall (callable=0x4addfb0, args=args@entry=0x7ffed5d5fd18, nargs=<optimized out>, keywords=<optimized out>, keywords@entry=0x0) at /usr/local/src/conda/python-3.8.20/Objects/call.c:159 #50 0x00000000004dcd27 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d5fd18, callable=<optimized out>) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:125 #51 _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d5fd18, callable=<optimized out>) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:115 #52 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #53 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3500 #54 0x00000000004e8486 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffed5d5fba0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #55 function_code_fastcall (globals=<optimized out>, nargs=1, args=<optimized out>, co=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:284 #56 _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ffed5d5fb78, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:411 #57 0x00000000004d8ac2 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d5fb78, callable=0x7fffea3b6040) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #58 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #59 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3486 #60 0x00000000004f5556 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffed5d5fa00) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #61 function_code_fastcall (globals=<optimized out>, nargs=1, args=<optimized out>, co=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:284 #62 _PyFunction_Vectorcall (kwnames=<optimized out>, nargsf=<optimized out>, stack=0x7ffed5d68cb0, func=0x7fffea3b6280) at /usr/local/src/conda/python-3.8.20/Objects/call.c:411 #63 _PyObject_Vectorcall (kwnames=<optimized out>, nargsf=<optimized out>, args=0x7ffed5d68cb0, callable=0x7fffea3b6280) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #64 method_vectorcall (method=<optimized out>, args=0x7ffed5d68cb8, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/classobject.c:60 #65 0x00000000004dcdc9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffed5d68cb8, callable=0x7ffeb9847980) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #66 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #67 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3469 #68 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffed5d68b20) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #69 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=locals@entry=0x0, args=args@entry=0x7ffff4961080, argcount=argcount@entry=2, kwnames=<optimized out>, kwargs=0x7ffff4961090, kwcount=<optimized out>, kwstep=1, defs=0x0, defcount=<optimized out>, kwdefs=0x0, closure=0x7ffeee72a310, name=0x7ffff7e1d530, qualname=0x7fffe4dd7850) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #70 0x00000000004e851c in _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ffff4961080, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:436 #71 0x00000000004e0882 in _PyObject_FastCallDict (callable=0x7ffeb91a2a60, args=0x7fffffffcaa0, nargsf=<optimized out>, kwargs=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:104 #72 0x00000000004f2143 in _PyObject_Call_Prepend (kwargs=0x7ffea49da240, args=0x7ffed5c690d0, obj=0x7ffed5d5d850, callable=0x7ffeb91a2a60) at /usr/local/src/conda/python-3.8.20/Objects/call.c:888 #73 slot_tp_init (self=0x7ffed5d5d850, args=0x7ffed5c690d0, kwds=0x7ffea49da240) at /usr/local/src/conda/python-3.8.20/Objects/typeobject.c:6790 #74 0x00000000004e10e3 in type_call (kwds=0x7ffea49da240, args=0x7ffed5c690d0, type=0x7182440) at /usr/local/src/conda/python-3.8.20/Objects/typeobject.c:994 #75 _PyObject_MakeTpCall (callable=callable@entry=0x7182440, args=<optimized out>, nargs=<optimized out>, keywords=<optimized out>, keywords@entry=0x7fffe4dd8350) at /usr/local/src/conda/python-3.8.20/Objects/call.c:159 #76 0x00000000004dd435 in _PyObject_Vectorcall (kwnames=0x7fffe4dd8350, nargsf=<optimized out>, args=<optimized out>, callable=0x7182440) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:125 #77 _PyObject_Vectorcall (kwnames=0x7fffe4dd8350, nargsf=<optimized out>, args=<optimized out>, callable=0x7182440) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:115 #78 call_function (kwnames=0x7fffe4dd8350, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #79 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3515 #80 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0xbc56ed0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #81 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=locals@entry=0x0, args=args@entry=0x7ae0d70, argcount=argcount@entry=5, kwnames=<optimized out>, kwargs=0x7ae0d98, kwcount=<optimized out>, kwstep=1, defs=0x7fffe4bb1208, defcount=<optimized out>, kwdefs=0x0, closure=0x0, name=0x7ffff7a6f300, qualname=0x7ffff7a6f300) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #82 0x00000000004e851c in _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ae0d70, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:436 #83 0x00000000004d9566 in _PyObject_Vectorcall (kwnames=0x7ffff7aca4c0, nargsf=<optimized out>, args=<optimized out>, callable=0x7ffeb91a28b0) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #84 call_function (kwnames=0x7ffff7aca4c0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #85 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3515 #86 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0x7ae0920) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #87 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=locals@entry=0x0, args=args@entry=0x7ac16a0, argcount=argcount@entry=4, kwnames=<optimized out>, kwargs=0x7ac16c0, kwcount=<optimized out>, kwstep=1, defs=0x0, defcount=<optimized out>, kwdefs=0x0, closure=0x0, name=0x7ffff78007b0, qualname=0x7ffff78007b0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #88 0x00000000004e851c in _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ac16a0, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:436 #89 0x00000000004d87a9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ac16a0, callable=0x7fffe5559820) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #90 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #91 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3500 #92 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0x7ac1320) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #93 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=locals@entry=0x0, args=args@entry=0x7ffff7b4d5b0, argcount=argcount@entry=1, kwnames=<optimized out>, kwargs=0x7ffff7b4d5b8, kwcount=<optimized out>, kwstep=1, defs=0x7ffea4e3ef28, defcount=<optimized out>, kwdefs=0x0, closure=0x0, name=0x7ffff7b7a8b0, qualname=0x7ffff7b7a8b0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #94 0x00000000004e851c in _PyFunction_Vectorcall (func=<optimized out>, stack=0x7ffff7b4d5b0, nargsf=<optimized out>, kwnames=<optimized out>) at /usr/local/src/conda/python-3.8.20/Objects/call.c:436 #95 0x00000000004d87a9 in _PyObject_Vectorcall (kwnames=0x0, nargsf=<optimized out>, args=0x7ffff7b4d5b0, callable=0x7ffea49d6700) at /usr/local/src/conda/python-3.8.20/Include/cpython/abstract.h:127 #96 call_function (kwnames=0x0, oparg=<optimized out>, pp_stack=<synthetic pointer>, tstate=0x743fe0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4963 #97 _PyEval_EvalFrameDefault (f=<optimized out>, throwflag=<optimized out>) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:3500 #98 0x00000000004d73d1 in PyEval_EvalFrameEx (throwflag=0, f=0x7ffff7b4d440) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:741 #99 _PyEval_EvalCodeWithName (_co=<optimized out>, globals=<optimized out>, locals=<optimized out>, args=<optimized out>, argcount=argcount@entry=0, kwnames=<optimized out>, kwargs=0x0, kwcount=<optimized out>, kwstep=2, defs=0x0, defcount=<optimized out>, kwdefs=0x0, closure=0x0, name=0x0, qualname=0x0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4298 #100 0x0000000000586079 in PyEval_EvalCodeEx (_co=_co@entry=0x7ffff7a39c90, globals=globals@entry=0x7ffff7b90240, locals=locals@entry=0x7ffff7b90240, args=args@entry=0x0, argcount=argcount@entry=0, kws=kws@entry=0x0, kwcount=0, defs=0x0, defcount=0, kwdefs=0x0, closure=0x0) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:4327 #101 0x000000000058603b in PyEval_EvalCode (co=co@entry=0x7ffff7a39c90, globals=globals@entry=0x7ffff7b90240, locals=locals@entry=0x7ffff7b90240) at /usr/local/src/conda/python-3.8.20/Python/ceval.c:718 #102 0x00000000005a5da1 in run_eval_code_obj (co=co@entry=0x7ffff7a39c90, globals=globals@entry=0x7ffff7b90240, locals=locals@entry=0x7ffff7b90240) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:1166 #103 0x00000000005a4daf in run_mod (mod=mod@entry=0x82fe98, filename=filename@entry=0x7ffff7aba5f0, globals=globals@entry=0x7ffff7b90240, locals=locals@entry=0x7ffff7b90240, flags=flags@entry=0x7fffffffd658, arena=arena@entry=0x7ffff7bc07f0) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:1188 #104 0x000000000045c624 in pyrun_file (fp=fp@entry=0x73f340, filename=filename@entry=0x7ffff7aba5f0, start=start@entry=257, globals=globals@entry=0x7ffff7b90240, locals=locals@entry=0x7ffff7b90240, closeit=closeit@entry=1, flags=0x7fffffffd658) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:1085 #105 0x000000000045c186 in pyrun_simple_file (flags=0x7fffffffd658, closeit=1, filename=0x7ffff7aba5f0, fp=0x73f340) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:439 #106 PyRun_SimpleFileExFlags (fp=fp@entry=0x73f340, filename=<optimized out>, closeit=closeit@entry=1, flags=flags@entry=0x7fffffffd658) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:472 #107 0x000000000045bdbd in PyRun_AnyFileExFlags (fp=fp@entry=0x73f340, filename=<optimized out>, closeit=closeit@entry=1, flags=flags@entry=0x7fffffffd658) at /usr/local/src/conda/python-3.8.20/Python/pythonrun.c:90 #108 0x000000000044fcae in pymain_run_file (cf=0x7fffffffd658, config=0x7433a0) at /usr/local/src/conda/python-3.8.20/Modules/main.c:391 #109 pymain_run_python (exitcode=0x7fffffffd650) at /usr/local/src/conda/python-3.8.20/Modules/main.c:620 #110 Py_RunMain () at /usr/local/src/conda/python-3.8.20/Modules/main.c:699 #111 0x000000000057a0d9 in Py_BytesMain (argc=<optimized out>, argv=<optimized out>) at /usr/local/src/conda/python-3.8.20/Modules/main.c:1131 #112 0x00007ffff7c2a1ca in __libc_start_call_main (main=main@entry=0x57a090 <main>, argc=argc@entry=16, argv=argv@entry=0x7fffffffd898) at ../sysdeps/nptl/libc_start_call_main.h:58 #113 0x00007ffff7c2a28b in __libc_start_main_impl (main=0x57a090 <main>, argc=16, argv=0x7fffffffd898, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffffffd888) at ../csu/libc-start.c:360 #114 0x0000000000579f8d in _start () ``` ### Environment - **OS:** Windows Subsystem for Linux (WSL2) - Ubuntu24.04 - **WSL** 2.6.3.0 - **kernel** - 6.6.87.2-microsoft-standard-WSL2 - **Python:** 3.8.20 - **PyTorch:** 2.2.2+cu121 ### Minimal Reproducible Example _No response_ ### Additional _No response_ ### Are you willing to submit a PR? - [ ] Yes I'd like to help by submitting a PR!