-
왜 ViT 가 잘 working 할까에 대해 연구한 논문.
[paper](https://arxiv.org/abs/2202.06709)
일반적으로 생각하는 MSA 가 좋은 이유
```
MSA 의 어떤 부분이 모델을 위해 좋을까?
==> long range dependency
MSA가 conv 처럼 동작할까?
==> MSA 가 general…
-
When indexing the parser tables with the next token (ignoring extended identifier behavior for this issue), we do one of two things:
* If the next token is an naked atom `s` (i.e. a symbol or keyword…
-
Port and cleanup the Lean 3 coinductive predicate meta program to Lean 4.
https://github.com/leanprover-community/mathlib/blob/76a3b82dc0ae6c9a60a8714d74c6cf2b854cf17b/src/meta/coinductive_predicates…
-
### Description of the problem
Consider this little example:
~~~coq
From Coq Require Import FinFun.
From Coq Require Import Classes.Equivalence.
Definition f_id {X: Type} (x0: X) := x0.
Lemm…
-
Writing the basics of polynomial rings, I am scratching my head about a good definition of such.
The first idea that comes to mind is to define R[X] as the sum over n:N of Rn[X], where the latter t…
-
@jozefg wrote some code on Grothendieck universes, available at https://github.com/unimath2019-tt/UniMath/pull/1.
Can it get merged into UniMath?
@mortberg
-
exists では通らず、use が通るようなコード例を探したい
-
Mathlib uses this idiom many times:
```lean
@[class] def ...
```
In Lean4, only structures can be classes. The porting tool ignores such class annotations. Ideally, Mathlib backports this restrict…
-
Note: the issue was created automatically with bugzilla2github tool
Original bug ID: BZ#5276
From: @xavierleroy
Reported version: 8.5
CC: @herbelin, @JasonGross
See also: [BZ#5277](https://github.…
-
We had an example where we need to assume a recursive function to be opaque while we verify it, but later we want to see its body to prove an inductive lemma. Thus, we want to control opaqueness at th…