Comments (7)
方便多贴一些代码和截图吗 这里面网络的结构不需要改变 只需要改变一下这里的张量形状 可以发一下你运行的代码脚本和相关参数设定
from unireplknet.
可以的,我使用的数据是按照readme里面所说的在点云data文件中下载,并放到了dataset文件夹中,运行结果如截图所示。
运行的项目为Time-Series,在该项目文件夹中所使用的unireplknet.py为从主文件夹复制,文件夹格式安排为:
运行方法:按照Time-Series中readme所提示的命令运行,即bash ./scripts/Global_Temp/UniRepLKNet.sh
,并无其他改动。
并没有使用iGEMM,使用默认pytorch卷积。
由于Time-Series中的models中不包含Corrformer_Large.py文件,所以在exp/exp_main.py文件中,在import中删除了Corrformer_Large,只保留from models import Corrformer
。
其余全部使用默认代码。
from unireplknet.
from unireplknet.
同学你好,你这个问题是因为你调用外层的UniRepLKNet的时候前传过了ImageNet的分类头所以特征被分类成了(1000种类别),但是在时序里面我们仅使用UniRepLKNet所提取的特征。
除此以外,可以对照一下我给出的对应的代码和张量形状输出
核心使用UniRepLKNet前传的代码如下面所示,对应github上面开源部分L126到L134
B, L, D = x_enc.shape
print((x_enc.shape))
_, _, C = x_mark_enc.shape
x_enc = x_enc.view(B, L, self.node_num, -1).permute(0, 2, 1, 3).contiguous() \
.view(B * self.node_num, L, D // self.node_num)
print((x_enc.shape))
x_mark_enc = x_mark_enc.unsqueeze(1).repeat(1, self.node_num, 1, 1).view(B * self.node_num, L, C)
print((x_enc.shape))
enc_out = self.enc_embedding(x_enc, x_mark_enc)
print((enc_out.shape))
enc_out = enc_out.reshape(self.node_num,1,192,192)
print((enc_out.shape))
from unireplknet.
谢谢,祝您早日发表!
from unireplknet.
replk/UniRepLKNet/Time-Series/layers/Multi_Correlation.py", line 111, in forward
_, S, _ = keys.shape
ValueError: too many values to unpack (expected 3)
根据上述更改后,会出现这个bug,请教下怎么改
from unireplknet.
replk/UniRepLKNet/Time-Series/layers/Multi_Correlation.py", line 111, in forward _, S, _ = keys.shape ValueError: too many values to unpack (expected 3) 根据上述更改后,会出现这个bug,请教下怎么改
是参数过多,需要去除一些
from unireplknet.
Related Issues (16)
- cp: 无法获取'ModelNet40' 的文件状态(stat): 没有那个文件或目录 HOT 1
- Layer Decay impl only supports convnext_small/base/large/xlarge HOT 1
- 图像分割
- input_size 384*128 tends to be nan HOT 3
- The training is very slow HOT 3
- 关于UniRepLKNetBlock中的self.gamma的作用
- Why UniRepLKNet does not using drop path.
- Install install 19_large_depthwise_conv2d_torch_extension but ”bash: ./setup.py: Permission denied“ HOT 1
- The parameter loading is incomplete HOT 2
- 论文里面的一个问题 HOT 4
- can it convert to onnx and trt to boost the inference speed? HOT 1
- error: command 'C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.4\\bin\\nvcc.exe' HOT 1
- depthwise conv 似乎没加速 HOT 1
- A small bug in reparameterize() HOT 2
- 您好,无法使用timm创建模型 HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from unireplknet.