Browse Source

check `batch_size % utilized_device_count` (#3276)

Bug fix to check batch_size divisibility of utilized CUDA device count vs total system CUDA device count.
modifyDataloader
Glenn Jocher GitHub 3 years ago
parent
commit
f3402353fb
No known key found for this signature in database GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 3 deletions
  1. +4
    -3
      utils/torch_utils.py

+ 4
- 3
utils/torch_utils.py View File

@@ -72,11 +72,12 @@ def select_device(device='', batch_size=None):

cuda = not cpu and torch.cuda.is_available()
if cuda:
n = torch.cuda.device_count()
if n > 1 and batch_size: # check that batch_size is compatible with device_count
devices = device.split(',') if device else range(torch.cuda.device_count()) # i.e. 0,1,6,7
n = len(devices) # device count
if n > 1 and batch_size: # check batch_size is divisible by device_count
assert batch_size % n == 0, f'batch-size {batch_size} not multiple of GPU count {n}'
space = ' ' * len(s)
for i, d in enumerate(device.split(',') if device else range(n)):
for i, d in enumerate(devices):
p = torch.cuda.get_device_properties(i)
s += f"{'' if i == 0 else space}CUDA:{d} ({p.name}, {p.total_memory / 1024 ** 2}MB)\n" # bytes to MB
else:

Loading…
Cancel
Save