Issue with Detectron2 and CUDA

[FrancoisPorcher]

Hi, congrats for the great work!

I am writing you because I am trying to evaluate EVA-02 on LVIS and I have problems with the Detectron2 Library:

• I tried to keep the detectron2 folder as you gave it, and I have PyTorch 2.0.0 and CUDA 11.8, and I get the following error:

Traceback (most recent call last):
File "/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/tools/lazyconfig_train_net.py", line 197, in
launch(
File "/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/detectron2/engine/launch.py", line 67, in launch
mp.spawn(
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 246, in spawn
return start_processes(fn, args, nprocs, join, daemon, start_method="spawn")
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 202, in start_processes
while not context.join():
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 163, in join
raise ProcessRaisedException(msg, error_index, failed_process.pid)
torch.multiprocessing.spawn.ProcessRaisedException:

-- Process 1 terminated with the following error:
Traceback (most recent call last):
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/multiprocessing/spawn.py", line 74, in _wrap
fn(i, *args)
File "/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/detectron2/engine/launch.py", line 124, in _distributed_worker
comm.synchronize()
File "/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/detectron2/utils/comm.py", line 82, in synchronize
dist.barrier(device_ids=[torch.cuda.current_device()])
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/distributed/c10d_logger.py", line 47, in wrapper
return func(*args, **kwargs)
File "/private/home/francoisporcher/anaconda3/envs/asuka/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 3696, in barrier
work = default_pg.barrier(opts=opts)
RuntimeError: CUDA error: the provided PTX was compiled with an unsupported toolchain.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

• I also tried to remove the detectron2 folder and clone the latest version because I thought it could be cause by CUDA incompatibility, but I have this error then:

Traceback (most recent call last):
File "/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/tools/lazyconfig_train_net.py", line 33, in
from detectron2.modeling import GeneralizedRCNNWithTTA, ema
ImportError: cannot import name 'ema' from 'detectron2.modeling' (/private/home/francoisporcher/WeakSupSeg/francois-semi-supervised/EVA/EVA-02/det/detectron2/detectron2/modeling/init.py)