Blealtan/efficient-kan issues and pull requests

#51 - Kolmogorov-Arnold Transformer

Issue - State: open - Opened by Adamdad 4 days ago

#50 - torch._C._LinAlgError: torch.linalg.lstsq: (Batch element 1): The least squares solution could not be computed because the input matrix does not have full rank (error code: 4).

Issue - State: open - Opened by black-yt 17 days ago - 2 comments

#49 - Add missing sp_trainable and sb_trainable params. Add Gaussian Pe…

Pull Request - State: open - Opened by MrPio about 2 months ago

#48 - Constant loss, network is not learning

Issue - State: closed - Opened by SuperKogito 3 months ago - 3 comments

#47 - How to cite

Issue - State: open - Opened by dgkoenigsberg 3 months ago - 1 comment

#46 - Questions

Issue - State: open - Opened by jindeok 3 months ago

#45 - flops and params

Issue - State: open - Opened by songxinghui 4 months ago - 1 comment

#44 - I would like to add a train test function to the KAN class

Issue - State: open - Opened by riteshshergill 4 months ago

#43 - which types of images could be better recognised using KAN approach?

Issue - State: open - Opened by matsevytyi 4 months ago

#42 - refactor: utilize Einops in appropriate areas to enhance readability

Pull Request - State: open - Opened by juvi21 4 months ago

#41 - block: [0,0,0], thread: [1,0,0] Assertion `input_val >= zero && input_val <= one` failed. when using BCEloss

Issue - State: open - Opened by Treakun 4 months ago

#40 - How to pip install?

Issue - State: closed - Opened by piwawa 4 months ago - 3 comments

#39 - ONNX

Issue - State: open - Opened by alonejxy 4 months ago

#38 - Something different from the official results for KAN.

Issue - State: open - Opened by ASCIIJK 4 months ago - 1 comment

#37 - Fix issue with view on previously transformed data

Pull Request - State: closed - Opened by akaashdash 4 months ago

#36 - Layer fails on inputs that have been transposed previously

Issue - State: closed - Opened by tRosenflanz 4 months ago - 1 comment

#30 - Train Multi-KANs simultaneously! Coding realization

Issue - State: closed - Opened by dwtum 4 months ago - 2 comments

#29 - Can still plot model and prune model in efficient kan?

Issue - State: open - Opened by yhaddbb 4 months ago - 12 comments

#28 - Can configure efficient-kan model for continual learning?

Issue - State: open - Opened by lukmanulhakeem97 4 months ago - 9 comments
Labels: enhancement

#27 - Symbolic Formulation from efficient-KAN

Issue - State: closed - Opened by MinSiThu 4 months ago - 1 comment

#26 - AssertionError

Issue - State: open - Opened by leedrake5 4 months ago - 1 comment

#25 - please report a bug to PyTorch. torch.linalg.lstsq: (Batch element 0): Argument 6 has illegal value. Most certainly there is a bug in the implementation calling the backend library.

Issue - State: closed - Opened by maomaocun 4 months ago - 1 comment

#24 - CUDA out of memory.

Issue - State: closed - Opened by ybu-lxd 4 months ago - 2 comments

#23 - Difference in memory usage of mlp and kan

Issue - State: open - Opened by c-pupil 4 months ago - 1 comment

#22 - RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [800, 256]], which is output 0 of AsStridedBackward0, is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

Issue - State: open - Opened by boostingboosting 4 months ago - 1 comment

#21 - Symbolic expression support

Issue - State: closed - Opened by maramos874 4 months ago - 1 comment

#20 - Intel oneMKL ERROR: Parameter 6 was incorrect on entry to SGELSY.

Issue - State: open - Opened by wza13 4 months ago - 7 comments

#19 - Supports inputs more than 2 dimensions For efficient-KAN

Issue - State: closed - Opened by WhatMelonGua 4 months ago - 5 comments

#18 - Are the "efficient-kan" and "official-kan" equivalent in terms of algorithms?

Issue - State: closed - Opened by yuedajiong 4 months ago - 12 comments
Labels: discussion

#17 - Is there anyway that I can save the model for later use

Issue - State: closed - Opened by binhnguyen7tiki 4 months ago - 7 comments

#16 - What is the mathematic equation behind the model?

Issue - State: closed - Opened by hoangthangta 4 months ago - 4 comments
Labels: discussion

#15 - KAN has more learnable parameters?

Issue - State: closed - Opened by ningwanyi 5 months ago - 2 comments

#14 - KAN intuition for setting the hidden layers

Issue - State: closed - Opened by marcopeix 5 months ago - 1 comment
Labels: discussion

#13 - Can MLP replace transformer?

Issue - State: closed - Opened by taotangSeek 5 months ago - 1 comment
Labels: discussion

#12 - Update test_simple_math.py

Pull Request - State: open - Opened by ODamilare 5 months ago

#11 - WIP: 2d convolution with KANs

Pull Request - State: closed - Opened by kabachuha 5 months ago - 3 comments

#10 - fix bug: false INTERNAL ASSERT FAILED

Pull Request - State: open - Opened by AmbroseX 5 months ago

#9 - Can anyone put kan(attention and mlp) into llama2.c?

Issue - State: closed - Opened by win10ogod 5 months ago - 1 comment

#8 - KANs are MLPs? MLP Equivalent

Issue - State: closed - Opened by hypnopump 5 months ago - 2 comments
Labels: discussion

GitHub / Blealtan/efficient-kan issues and pull requests