Pytorch 2.4.1 compilation on jetson xavier nx

Robotics & Edge Computing Jetson & Embedded Systems Jetson AGX Xavier
, ,
1

Compiled successfully on my jetson xavier nx 8GB

but now when I uses it with coqui-tts I get:

NvMapMemAllocInternalTagged: 1074810371 error 12
NvMapMemHandleAlloc: error 0
NvMapMemAllocInternalTagged: 1074810371 error 12
NvMapMemHandleAlloc: error 0
0.00%: 0/10
CUDA error: out of memory
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Traceback (most recent call last):
File “/home/workmin/repos/ebook2audiobook/lib/classes/tts_engines/yourtts.py”, line 76, in convert
self.engine.to(device)
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 1174, in to
return self._apply(convert)
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 780, in _apply
module._apply(fn)
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 780, in _apply
module._apply(fn)
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 780, in _apply
module._apply(fn)
[Previous line repeated 5 more times]
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 805, in _apply
param_applied = fn(param)
File “/home/workmin/repos/ebook2audiobook/python_env/lib/python3.10/site-packages/torch/nn/modules/module.py”, line 1160, in convert
return t.to(
RuntimeError: CUDA error: out of memory
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File “/home/workmin/repos/ebook2audiobook/lib/core.py”, line 1689, in convert_chapters2audio
success = tts_manager.convert_sentence2audio(sentence_number, sentence) if sentence else True
File “/home/workmin/repos/ebook2audiobook/lib/classes/tts_manager.py”, line 23, in convert_sentence2audio
return self.engine.convert(sentence_number, sentence)
File “/home/workmin/repos/ebook2audiobook/lib/classes/tts_engines/yourtts.py”, line 153, in convert
raise ValueError(e)
ValueError: CUDA error: out of memory
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

Caught DependencyError: CUDA error: out of memory
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

any idea?

3

Hi,

The compiler reports an out-of-memory error, which might be related to the following issue:

You can try to reflash your environment back to r35.6.2 to avoid this issue.
Thanks.

4

the flash is new, booting on mmc and rootfs on nvme…

5

Hi

The issue happens on r35.6.3.
You can check your environment with the command below:

$ cat /etc/nv_tegra_release

Thanks.

6

yes I know that’s why I opened this topic :)

R35 (release), REVISION: 6.3, GCID: 12345678, BOARD: t186ref, EABI: aarch64, DATE: Tue 28 Oct 2025 12:57:59 AM UTC

7

It would be so nice to let xavier nx users to upgrade to jetpack 6+ so I will be able to use python 3.10+ at least! let’s protect the planet and keep electronic 10 years at least!

8

Hi,

We are still preparing the changes for r35.6.3.
Currently, you can use r35.6.2 to avoid this issue.

Thanks.

9

I prefer to wait your fixes as I have no more time to pass to downgrade.