Super-gradients version 3.1.1 is giving import error

[PrajwalCogniac]

🐛 Describe the bug

Super-gradients version 3.1.1 is giving import error

If you try running the ipython notebooks for yolo nas fine-tuning in google collab, it was working fine till 2 days back now once the pip installations (super-gradients==3.1.1) are done and after you restart the runtime you get import error.

But if you install the latest version of the super-gradients, and then run the following block runs but we get error in this code-block

Versions

Its standard google colab

--2023-07-18 18:54:23-- https://raw.githubusercontent.com/pytorch/pytorch/main/torch/utils/collect_env.py
Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...
Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 21653 (21K) [text/plain]
Saving to: ‘collect_env.py’

collect_env.py 100%[===================>] 21.15K --.-KB/s in 0.001s

2023-07-18 18:54:24 (24.6 MB/s) - ‘collect_env.py’ saved [21653/21653]

Collecting environment information...
PyTorch version: 2.0.1+cu118
Is debug build: False
CUDA used to build PyTorch: 11.8
ROCM used to build PyTorch: N/A

OS: Ubuntu 20.04.6 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0
Clang version: 10.0.0-4ubuntu1
CMake version: version 3.25.2
Libc version: glibc-2.31

Python version: 3.10.12 (main, Jun 7 2023, 12:45:35) [GCC 9.4.0] (64-bit runtime)
Python platform: Linux-5.15.109+-x86_64-with-glibc2.31
Is CUDA available: True
CUDA runtime version: 11.8.89
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: Tesla T4
Nvidia driver version: 525.105.17
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.0
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.0
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 2
On-line CPU(s) list: 0,1
Thread(s) per core: 2
Core(s) per socket: 1
Socket(s): 1
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) CPU @ 2.00GHz
Stepping: 3
CPU MHz: 2000.208
BogoMIPS: 4000.41
Hypervisor vendor: KVM
Virtualization type: full
L1d cache: 32 KiB
L1i cache: 32 KiB
L2 cache: 1 MiB
L3 cache: 38.5 MiB
NUMA node0 CPU(s): 0,1
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Mitigation; PTE Inversion
Vulnerability Mds: Vulnerable; SMT Host state unknown
Vulnerability Meltdown: Vulnerable
Vulnerability Mmio stale data: Vulnerable
Vulnerability Retbleed: Vulnerable
Vulnerability Spec store bypass: Vulnerable
Vulnerability Spectre v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers
Vulnerability Spectre v2: Vulnerable, IBPB: disabled, STIBP: disabled, PBRSB-eIBRS: Not affected
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Vulnerable
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc rep_good nopl xtopology nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap clflushopt clwb avx512cd avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves arat md_clear arch_capabilities

Versions of relevant libraries:
[pip3] numpy==1.22.4
[pip3] torch==2.0.1+cu118
[pip3] torchaudio==2.0.2+cu118
[pip3] torchdata==0.6.1
[pip3] torchsummary==1.5.1
[pip3] torchtext==0.15.2
[pip3] torchvision==0.15.2+cu118
[pip3] triton==2.0.0
[conda] Could not collect

[kenhuang1964]

I'm getting the same thing on colab. Need help as well.

[zama-sarib]

Use this command to install super_gradient.
pip install git+https://github.com/Deci-AI/super-gradients.git.
Working fine on colab for me.

[PrajwalCogniac]

Thanks a lot for the reply @zama-sarib
I know the command above does the trick, that was even mentioned in the Quantization Aware Training YoloNAS on Custom Dataset Notebook, but it does install the latest version of 3.1.3 and that works correctly.
However the earlier version of 3.1.1 gives you this import error which is strange.

[mwyborski]

Doesn't work either for me. Fails directly at import super_gradients.

[23pointsNorth]

I now install 3.1.3 and have to change a few of the code boxes below for the new api - e.g.

from super_gradients import Trainer, setup_device
from super_gradients.training import  MultiGPUMode

CHECKPOINT_DIR = '/home/notebook_ckpts/'
setup_device(device="cuda", multi_gpu=MultiGPUMode.OFF, num_gpus=1)
trainer = Trainer(experiment_name="segmentation_quick_start", ckpt_root_dir=CHECKPOINT_DIR)

[Louis-Dupont]

This happened due to a breaking change introduced in python>=3.10.
It looks like google collab upgraded the default version to python~=3.10 which explains why it used to work but then stopped working without any apparent reason.

We now fixed this issue and it is part of the latest release super-gradients==3.1.3

[23pointsNorth]

@Louis-Dupont any chance the colab example notebooks are updated for 3.1.3 as well?

[BloodAxe]

3.1.3 is on PyPy already, you're welcome to update your notebooks.
Hopefully this upgrade should resolve your issue. I'm closing it rn as we were able to verify it fixes the import problem locally. If you still experience this - feel free to re-open.