File size: 1,361 Bytes
985cc7f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 |
import pytest
import torch
from mmseg.core import OHEMPixelSampler
from mmseg.models.decode_heads import FCNHead
def _context_for_ohem():
return FCNHead(in_channels=32, channels=16, num_classes=19)
def test_ohem_sampler():
with pytest.raises(AssertionError):
# seg_logit and seg_label must be of the same size
sampler = OHEMPixelSampler(context=_context_for_ohem())
seg_logit = torch.randn(1, 19, 45, 45)
seg_label = torch.randint(0, 19, size=(1, 1, 89, 89))
sampler.sample(seg_logit, seg_label)
# test with thresh
sampler = OHEMPixelSampler(
context=_context_for_ohem(), thresh=0.7, min_kept=200)
seg_logit = torch.randn(1, 19, 45, 45)
seg_label = torch.randint(0, 19, size=(1, 1, 45, 45))
seg_weight = sampler.sample(seg_logit, seg_label)
assert seg_weight.shape[0] == seg_logit.shape[0]
assert seg_weight.shape[1:] == seg_logit.shape[2:]
assert seg_weight.sum() > 200
# test w.o thresh
sampler = OHEMPixelSampler(context=_context_for_ohem(), min_kept=200)
seg_logit = torch.randn(1, 19, 45, 45)
seg_label = torch.randint(0, 19, size=(1, 1, 45, 45))
seg_weight = sampler.sample(seg_logit, seg_label)
assert seg_weight.shape[0] == seg_logit.shape[0]
assert seg_weight.shape[1:] == seg_logit.shape[2:]
assert seg_weight.sum() == 200
|