orientLoss
orientLoss is a classification loss function. It less prone to overfitting compared to crossEntropy. Compatible with random vector/embeddings with direction-indicating categories、probabilistic、oneHot as target.
思路说明
很多时候网络需要输出的类别数远高于网络中使用的特征数/通道数,需要在最后一层加一个巨大的线性层将输出扩增才送入交叉熵损失.但本质上这一层并没有增加信息量,所以可以使用某些损失函数取代这一过程.再考虑到交叉熵由于其下降没有终点,容易引起过拟合.通过设计公式从形状上逼近和优化维度无限扩增后的交叉熵损失得到了这个损失函数.
Install
pip install orientLoss
Use
from orientLoss import orientLoss
...
loss=orientLoss(input,target,dim=-1,meanOut=True,angleSmooth=1,normSmooth=1,dimScalingOrd=0,eps=1e-8)
target可以是由方向指示类别的随机向量/嵌入、概率、独热编码. 概率或独热编码输入时最好各类别独立进行标准化后再送入损失函数.
1 maintainer
Metadata
1 maintainer
Release files for orientLoss 0.0.8
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| orientloss-0.0.8.tar.gz | 8.2 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| orientloss-0.0.8-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 16.6 kB
Release files / orientloss-0.0.8.tar.gz
| Download URL | orientloss-0.0.8.tar.gz |
|---|---|
| Size | 8.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
8b2760ce01a70dac8266ffa185b59387ca33d405d3edf018bdc549943aeadf0e
|
|
BLAKE2b-256 checksum How to use checksums |
01e67758be0d85352b8ccd9f9808ace74d8a308ca488283325f42c539c5bb0c2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.12.9
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 17, 2025.
Transparency logRelease files / orientloss-0.0.8-py3-none-any.whl
| Download URL | orientloss-0.0.8-py3-none-any.whl |
|---|---|
| Size | 8.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
99c2acc7a33d4aebf81ec4008403582d9db05c4d43755064d80ce29d8a14c28a
|
|
BLAKE2b-256 checksum How to use checksums |
56272a8a0ca2627ebdfc46dd8e5f3f0a99bd25b3d8dbe1106afd380d39af9d7d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.12.9
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 17, 2025.
Transparency log
