RuntimeError: Arguments for call are not valid

I was trying to test the repo “IsaacGymEnvs-main” with command: python train.py task=Carpole, an error occurs:

Importing module ‘gym_36’ (/home/eowyn/文档/IsaacGym_Preview_3_Package/isaacgym/python/isaacgym/_bindings/linux-x86_64/gym_36.so)
Setting GYM_USD_PLUG_INFO_PATH to /home/eowyn/文档/IsaacGym_Preview_3_Package/isaacgym/python/isaacgym/_bindings/linux-x86_64/usd/plugInfo.json
/home/eowyn/anaconda3/envs/sim_env/lib/python3.6/site-packages/torch/utils/cpp_extension.py:4: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module’s documentation for alternative uses
import imp
PyTorch version 1.4.0
Device count 0
/home/eowyn/文档/IsaacGym_Preview_3_Package/isaacgym/python/isaacgym/_bindings/src/gymtorch
Using /tmp/torch_extensions as PyTorch extensions root…
Emitting ninja build file /tmp/torch_extensions/gymtorch/build.ninja…
Building extension module gymtorch…
ninja: no work to do.
Loading extension module gymtorch…
Traceback (most recent call last):
File “train.py”, line 40, in
from isaacgymenvs.utils.rlgames_utils import RLGPUEnv, RLGPUAlgoObserver, get_rlgames_env_creator
File “/home/eowyn/isaacgym/IsaacGymEnvs-main/isaacgymenvs/utils/rlgames_utils.py”, line 36, in
from tasks import isaacgym_task_map
File “/home/eowyn/isaacgym/IsaacGymEnvs-main/isaacgymenvs/tasks/init.py”, line 31, in
from tasks.ant import Ant
File “/home/eowyn/isaacgym/IsaacGymEnvs-main/isaacgymenvs/tasks/ant.py”, line 37, in
from isaacgymenvs.utils.torch_jit_utils import *
File “/home/eowyn/isaacgym/IsaacGymEnvs-main/isaacgymenvs/utils/torch_jit_utils.py”, line 272, in
@torch.jit.script
File “/home/eowyn/anaconda3/envs/sim_env/lib/python3.6/site-packages/torch/jit/init.py”, line 1281, in script
fn = torch._C._jit_script_compile(qualified_name, ast, _rcb, get_default_args(obj))
RuntimeError:
Arguments for call are not valid.
The following variants are available:

aten::norm.Scalar(Tensor self, Scalar p=2) → (Tensor):
Keyword argument dim unknown.

aten::norm.ScalarOpt_dtype(Tensor self, Scalar? p, *, int dtype) → (Tensor):
Argument p not provided.

aten::norm.ScalarOpt_dim(Tensor self, Scalar? p, int[1] dim, bool keepdim=False) → (Tensor):
Argument p not provided.

aten::norm.ScalarOpt_dim_dtype(Tensor self, Scalar? p, int[1] dim, bool keepdim, *, int dtype) → (Tensor):
Argument p not provided.

aten::norm.out(Tensor self, Scalar? p, int[1] dim, bool keepdim=False, *, Tensor(a!) out) → (Tensor(a!)):
Argument p not provided.

aten::norm.dtype_out(Tensor self, Scalar? p, int[1] dim, bool keepdim, *, int dtype, Tensor(a!) out) → (Tensor(a!)):
Argument p not provided.

The original call is:
File “/home/eowyn/isaacgym/IsaacGymEnvs-main/isaacgymenvs/utils/torch_jit_utils.py”, line 276
min_theta = 1e-5

angle = torch.norm(exp_map, dim=-1)
        ~~~~~~~~~~ <--- HERE
angle_exp = torch.unsqueeze(angle, dim=-1)
axis = exp_map / angle_exp

Any advice? Thanks!

Hi @Eowyn

It seems that the problem is the torch version.
Installing a version higher than 1.7.0 can solve the problem.

1 Like

Thank you so much, problem solved~ I installed a wrong version pytorch, and it doesn’t support gpu.