problem with transcoder

Read TransCoder/translate.py around lines 83 and 171, then XLM/src/model/init.py around line 181 where model loading reaches CUDA initialization. Reproduce the reported Translator(params) startup in the shown Python 3.7 environment and inspect the CUDA failure. Done means the cause is identified and a verified resolution or precise environment requirement is recorded.

showing below error
File "TransCoder/translate.py", line 171, in
translator = Translator(params)
File "TransCoder/translate.py", line 83, in init
encoder, decoder = build_model(self.reloaded_params, self.dico)
File "/content/TransCoder/XLM/src/model/init.py", line 181, in build_model
enc_path, map_location=lambda storage, loc: storage.cuda(params.local_rank))
File "/usr/local/lib/python3.7/dist-packages/torch/serialization.py", line 608, in load
return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
File "/usr/local/lib/python3.7/dist-packages/torch/serialization.py", line 787, in _legacy_load
result = unpickler.load()
File "/usr/local/lib/python3.7/dist-packages/torch/serialization.py", line 743, in persistent_load
deserialized_objects[root_key] = restore_location(obj, location)
File "/usr/local/lib/python3.7/dist-packages/torch/serialization.py", line 830, in restore_location
result = map_location(storage, location)
File "/content/TransCoder/XLM/src/model/init.py", line 181, in
enc_path, map_location=lambda storage, loc: storage.cuda(params.local_rank))
File "/usr/local/lib/python3.7/dist-packages/torch/_utils.py", line 71, in _cuda
with torch.cuda.device(device):
File "/usr/local/lib/python3.7/dist-packages/torch/cuda/init.py", line 225, in enter
self.prev_idx = torch.cuda.current_device()
File "/usr/local/lib/python3.7/dist-packages/torch/cuda/init.py", line 432, in current_device
_lazy_init()
File "/usr/local/lib/python3.7/dist-packages/torch/cuda/init.py", line 172, in _lazy_init
torch._C._cuda_init()