AI Rank
EN
← 开源
raminmh

CfC

Closed-form Continuous-time Neural Networks

Model DevelopmentArchitecturePython
在 GitHub 打开
增长势头
+-124 小时新增 Star-0.1%
1.06k
Star
162
Fork
+1
本周
4
贡献者
创建于 2021-01-09 · 更新于 2026-10-16 · 今日第 18968 名
主要开发者
raminmhmlech26lcclauss

raminmh/CfC 排名第 18,968，有 1,056 个星标。

同类别比较

raminmh/CfC 排名第 18,968，此处有 8 个同类条目；drprojects/superpoint_transformer 排名第 18,964。

同类别比较
名称排名指标
openai/automated-interpretability18,957
语言:
Python
星标数:
1,082
分叉数:
132
rixwew/pytorch-fm18,960
语言:
Python
星标数:
1,080
分叉数:
228
简介:
Factorization Machine models in PyTorch
hukenovs/hagrid18,961
语言:
Python
星标数:
1,077
分叉数:
149
简介:
HAnd Gesture Recognition Image Dataset
drprojects/superpoint_transformer18,964
语言:
Python
星标数:
1,063
分叉数:
136
简介:
Official PyTorch implementation of Superpoint Transformer [ICCV'23], SuperCluster [3DV'24 Oral], and EZ-SP [ICRA'26]
ArrowLuo/CLIP4Clip18,974
语言:
Python
星标数:
1,033
分叉数:
138
简介:
An official implementation for "CLIP4Clip: An Empirical Study of CLIP for End to End Video Clip Retrieval"
alex-petrenko/sample-factory18,978
语言:
Python
星标数:
1,026
分叉数:
155
简介:
High throughput synchronous and asynchronous reinforcement learning
aiming-lab/SkillRL18,983
语言:
Python
星标数:
1,007
分叉数:
80
简介:
[NeurIPS'26] SkillRL: Evolving Agents via Recursive Skill-Augmented Reinforcement Learning
kang205/SASRec18,984
语言:
Python
星标数:
1,007
分叉数:
194
简介:
SASRec: Self-Attentive Sequential Recommendation
关键参数与规格
关键参数与规格
参数值
语言Python
类别Model Development
子类别Architecture
星标数1,056
分叉数162
贡献者4
创建日期2021-01-09
更新日期2026-10-16
1 天星标数-1
7 天星标数1
排名18,968
上次排名11,764
排名变化-7,204
7 天星标排名10,245
上次 7 天星标排名8,340
7 天星标排名变化-1,905
总星标排名10,712
上次总星标排名10,683
总星标排名变化-29
国家USA
README 摘录

Closed-form Continuous-time Models Closed-form Continuous-time Neural Networks (CfCs) are powerful sequential liquid neural information processing units. Paper Open Access: Arxiv: A Tutorial on Liquid Neural Networks including Liquid CfCs: Requirements Python3. 2 or newer Module description Implementation of the CfC (various versions) in Tensorflow 2. x Implementation of the CfC (various versions) in PyTorch Trains the CfC models on the Physionet 2012 dataset in PyTorch (code adapted from Rubanova et al. 2019) Trains the CfC models

on the XOR dataset in Tensorflow (code adapted from Lechner & Hasani, 2020) Trains the CfC models on the IMDB dataset in Tensorflow (code adapted from Keras examples website) Trains the CfC models on the Walker2d dataset in Tensorflow (code adapted from Lechner & Hasani, 2020) Datasets (same splits) from Lechner & Hasani (2020) and Physionet dataset (same split) from Rubanova et al. (2019) Usage All training scripts except the following three flags Runs the CfC without the (1-sigmoid) part Runs the

CfC direct solution Runs an LTC with a semi-implicit ODE solver instead of a CfC Mixes the CfC's RNN-state with a LSTM to avoid vanishing gradients If none of these flags are provided, the full CfC model is used For instance train the full CfC model on the Physionet dataset. Similarly runs the direct CfC solution on the walker2d dataset. For downloading the Walker2d dataset of Lechner & Hasani 2020, run Cite

常见问题

记录了多少星标和分叉？

raminmh/CfC 有 1,056 个星标和 162 个分叉。

仓库排名是多少？

raminmh/CfC 排名第 18,968。

记录的语言和类别是什么？

raminmh/CfC 有 1,056 个星标，语言为 Python，类别为 Model Development。

哪个仓库排名最接近？

raminmh/CfC 与 drprojects/superpoint_transformer 比较，后者排名第 18,964，有 1,063 个星标。

README

Closed-form Continuous-time Models

Closed-form Continuous-time Neural Networks (CfCs) are powerful sequential liquid neural information processing units.

Paper Open Access: https://www.nature.com/articles/s42256-022-00556-7

Arxiv: https://arxiv.org/abs/2106.13898

A Tutorial on Liquid Neural Networks including Liquid CfCs: https://ncps.readthedocs.io/en/latest/quickstart.html

Requirements

  • Python3.6 or newer
  • Tensorflow 2.4 or newer
  • PyTorch 1.8 or newer
  • pytorch-lightning 1.3.0 or newer
  • scikit-learn 0.24.2 or newer

Module description

  • tf_cfc.py Implementation of the CfC (various versions) in Tensorflow 2.x
  • torch_cfc.py Implementation of the CfC (various versions) in PyTorch
  • train_physio.py Trains the CfC models on the Physionet 2012 dataset in PyTorch (code adapted from Rubanova et al. 2019)
  • train_xor.py Trains the CfC models on the XOR dataset in Tensorflow (code adapted from Lechner & Hasani, 2020)
  • train_imdb.py Trains the CfC models on the IMDB dataset in Tensorflow (code adapted from Keras examples website)
  • train_walker.py Trains the CfC models on the Walker2d dataset in Tensorflow (code adapted from Lechner & Hasani, 2020)
  • irregular_sampled_datasets.py Datasets (same splits) from Lechner & Hasani (2020)
  • duv_physionet.py and duv_utils.py Physionet dataset (same split) from Rubanova et al. (2019)

Usage

All training scripts except the following three flags

  • no_gate Runs the CfC without the (1-sigmoid) part
  • minimal Runs the CfC direct solution
  • use_ltc Runs an LTC with a semi-implicit ODE solver instead of a CfC
  • use_mixed Mixes the CfC's RNN-state with a LSTM to avoid vanishing gradients

If none of these flags are provided, the full CfC model is used

For instance

python3 train_physio.py

train the full CfC model on the Physionet dataset.

Similarly

train_walker.py --minimal

runs the direct CfC solution on the walker2d dataset.

For downloading the Walker2d dataset of Lechner & Hasani 2020, run

source download_dataset.sh

Cite

@article{hasani_closed-form_2022,
	title = {Closed-form continuous-time neural networks},
	journal = {Nature Machine Intelligence},
	author = {Hasani, Ramin and Lechner, Mathias and Amini, Alexander and Liebenwein, Lucas and Ray, Aaron and Tschaikowski, Max and Teschl, Gerald and Rus, Daniela},
	issn = {2522-5839},
	month = nov,
	year = {2022},
}
排名相邻
同类项目