Nano_llm: libnvdla_compiler.so: cannot open shared object file

Had my Orin Nano super developer kit for a few days, I’m trying to run a very basic demo from the nano_vlm tutorial but the command crashes:

jetson-containers run $(autotag nano_llm) \
  python3 -m nano_llm.chat --api=mlc \
    --model Efficient-Large-Model/VILA1.5-3b \
    --max-context-len 256 \
    --max-new-tokens 32
...
ImportError: libnvdla_compiler.so: cannot open shared object file: No such file or directory

Reading around on this forum, I see that this file should be mounted from the host, however I don’t see that file:

# Don't think these results are relevant - just things from docker images
sudo find / -name libnvdla_compiler.so
/var/lib/docker/overlay2/4ddba773fe225e287fdaa8e94a72d5afe6cfb92a9e5c18ca681145ca70696d5d/diff/usr/lib/aarch64-linux-gnu/nvidia/libnvdla_compiler.so
/var/lib/docker/overlay2/6a81d8c11432e4070465681727e56362bebc4932f53873a8ae10e60813c4aa43/diff/usr/lib/aarch64-linux-gnu/nvidia/libnvdla_compiler.so

I tried sudo apt install cuda but still no joy. Here’s some info on my setup:

robin@orin:~$ jetson_release 
Software part of jetson-stats 4.3.2 - (c) 2024, Raffaello Bonghi
Model: NVIDIA Jetson Orin Nano Engineering Reference Developer Kit Super - Jetpack 6.2 [L4T 36.4.3]
NV Power Mode[2]: MAXN_SUPER
Serial Number: [XXX Show with: jetson_release -s XXX]
Hardware:
 - P-Number: p3767-0005
 - Module: NVIDIA Jetson Orin Nano (Developer kit)
Platform:
 - Distribution: Ubuntu 22.04 Jammy Jellyfish
 - Release: 5.15.148-tegra
jtop:
 - Version: 4.3.2
 - Service: Active
Libraries:
 - CUDA: 12.6.68
 - cuDNN: Not installed
 - TensorRT: Not installed
 - VPI: Not installed
 - Vulkan: 1.3.204
 - OpenCV: 4.5.4 - with CUDA: NO

Is there a way to install the missing files? I’d like to avoid re-flashing if poss.

Hi,

It looks like the file doesn’t exist in your environment.
Please try to install it and test it again:

$ apt-get update
$ apt-get install nvidia-l4t-dla-compiler

Thanks.