HongChuang/ultralytics-ascend
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 3

Update MLP module for RTDETR backward compatibility (#14901)

Browse source 
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
This commit is contained in:
Laughing 2024-08-04 02:03:33 +08:00 • committed by GitHub
parent 121f2242e1
commit 08263f5737
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 5 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

3
tests/test_cli.py
View file

@ -8,6 +8,7 @@ from PIL import Image
from tests import CUDA_DEVICE_COUNT, CUDA_IS_AVAILABLE
from ultralytics.cfg import TASK2DATA, TASK2MODEL, TASKS
from ultralytics.utils import ASSETS, WEIGHTS_DIR, checks
from ultralytics.utils.torch_utils import TORCH_1_9
# Constants
TASK_MODEL_DATA = [(task, WEIGHTS_DIR / TASK2MODEL[task], TASK2DATA[task]) for task in TASKS]
@ -57,6 +58,8 @@ def test_rtdetr(task="detect", model="yolov8n-rtdetr.yaml", data="coco8.yaml"):
# Warning: must use imgsz=640 (note also add coma, spaces, fraction=0.25 args to test single-image training)
run(f"yolo train {task} model={model} data={data} --imgsz= 160 epochs =1, cache = disk fraction=0.25")
run(f"yolo predict {task} model={model} source={ASSETS / 'bus.jpg'} imgsz=160 save save_crop save_txt")
if TORCH_1_9:
run(f"yolo predict {task} model='rtdetr-l.pt' source={ASSETS / 'bus.jpg'} imgsz=160 save save_crop save_txt")
@pytest.mark.skipif(checks.IS_PYTHON_3_12, reason="MobileSAM with CLIP is not supported in Python 3.12")

4
ultralytics/nn/modules/transformer.py
View file

@ -186,8 +186,8 @@ class MLP(nn.Module):
def forward(self, x):
"""Forward pass for the entire MLP."""
for i, layer in enumerate(self.layers):
x = self.act(layer(x)) if i < self.num_layers - 1 else layer(x)
return x.sigmoid() if self.sigmoid else x
x = getattr(self, "act", nn.ReLU())(layer(x)) if i < self.num_layers - 1 else layer(x)
return x.sigmoid() if getattr(self, "sigmoid", False) else x
class LayerNorm2d(nn.Module):