-
I was wondering if this could support ddrnet23 (not slim) as well
-
Line 282-283, models/others/ddrnet23_adb_bag.py
#self.bag = model_utils.BagFM(planes * 4, planes * 2, planes * 4)
self.dfm = model_utils.DFM3(planes * 4, planes * 4)
Neither of …
-
Thank you for the repo. Unfortunately, I was not able to reproduce the DDRNet-39 performance on cityscapes twice with default settings.
Is it possible to provide the weights for DDRNet-39 model on c…
-
### Describe the bug
torch.fx.proxy.TraceError: class `MMArchitectureQuant` in mmrazor/models/algorithms/quantization/mm_architecture.py: Proxy object cannot be iterated. This can be attempted when…
-
你好,你这代码写的很好鲁棒性很强!但是我在本地单卡3060训练DDRNet23_slim,训练参数如下的情况下,跑出来结果只有70%
```
python -m torch.distributed.launch --nproc_per_node=1 \
train.py --model DDRNet --out_stride 8 \
…
-
在我电脑上测了下 1650
STDC1-Seg50 一张速度23.1ms
STDC2-Seg50 28ms
DDRNET 速度一张11.27ms
(都是python下直接测的,我想转成tensorRT 降精度 时间缩短比例应该一样)
论文给出的MIOU STDC1低于 DDRNET
我觉得STDC整体表现是不如ddrnet-slim-23的
(我感觉不少人会想到看到直接…
-
如果使用ddrnet.pytorch里面,转成onnx, 再用trtexec 转换成engine文件,推理就可以吧?
-
在default_runtime.py中设置了可视化的参数,如下:
vis_backends = [dict(type='LocalVisBackend')]
visualizer = dict(
type='SegLocalVisualizer', vis_backends=vis_backends, name='visualizer')
在debug时,代码并没有在mmse…
-
The download links of DDRNet's Backbone and Weight are invalid
-
**Describe the issue**:
I compress the model: https://github.com/chenjun2hao/DDRNet.pytorch
when add this code "pruner._unwrap_model()", meet the error
Traceback (most recent call last):
Fil…