profile
viewpoint

dselsam/certigrad 374

Bug-free machine learning on stochastic computation graphs

dselsam/neurosat 208

NeuroSAT: Learning a SAT Solver from Single-Bit Supervision

dselsam/neurocore-public 29

NeuroCore: Guiding CDCL with Unsat-Core Predictions

dselsam/binport 10

A tool for building Lean4 .olean files from Lean3 export data

dselsam/lean-python-bindings 8

Python Bindings to the Lean Theorem Prover http://leanprover.github.io/

dselsam/arc 5

Experimental solver for ARC

dselsam/islanders 2

Formalization of the infamous blue-eyed islanders puzzle in Lean.

dselsam/lean 2

Lean Theorem Prover

dselsam/neurocuber 2

NeuroCuber: training NeuroSAT to make cubing decisions for hard SAT problems

issue commentleanprover-community/mathport

simp can't prove true

Yes, the actual problem is in trying to synthesize the missing arguments. One of the arguments is a Prop but the type has metavariables, and simp goes astray trying to discharge this mvar-filled goal. I think there should just be a no-mvar check before the discharge call here: https://github.com/leanprover/lean4/blob/master/src/Lean/Meta/Tactic/Simp/Rewrite.lean#L18-L20 @leodemoura

gebner

comment created time in 2 days

issue commentleanprover-community/mathport

simp can't prove true

Apparently the LHS of

Subtype.coe_mk : ∀ {α : Sort u_1} {p : α → Prop} (a : α) (h : p a), { val := a, property := h }.val = a

unifies with True, so rewriting with this lemma introduces a bunch of metavariables into the goal.

gebner

comment created time in 2 days

issue openedleanprover-community/mathport

`hasOfNatOfCoe` interacts badly with `AddUnits.hasCoe`

import Mathbin.Data.Real.Basic
#check (1 : ℝ)

times out with relevant trace:

⊢ OfNat ℝ 1
> apply hasOfNatOfCoe 
⊢ Coe ?x ℝ 
> apply AddUnits.hasCoe 
⊢ Coe (AddUnits ?x) ℝ
> apply AddUnits.hasCoe
⊢ Coe (AddUnits (AddUnits ?x)) ℝ
...

where the relevant types are:

AddUnits.hasCoe : {α : Type u_1} → [_inst_1 : AddMonoidₓ α] → Coe (AddUnits α) α
hasOfNatOfCoe : {α : Type u_1} → {β : Type u_2} → {n : ℕ} → [inst : Coe α β] → [inst : OfNat α n] → OfNat β n

Note that hasOfNatOfCoe will create a goal with a metavariable no matter the order the arguments are tried in.

created time in 2 days

issue openedleanprover/lake

Duplicate symbols with diamond dependencies

Suppose package sink depends on child1 and child2, both of which depend on root. Then (assuming some number of actual imports) lake will pass the root files twice to leanc. A self-contained repro is attached. When trying to build sink, the following leanc command will execute:

> leanc -o ./build/bin/sink ./build/ir/Main.o ./build/ir/Sink.o ./../child1/./build/ir/Child1.o ./../root/./build/ir/Root.o ./../root/./build/ir/Root.o

which (depending on the compiler/linker) may give a "duplicate symbol" error, e.g.:

error: ld64.lld: error: duplicate symbol: _initialize_Root
>>> defined in ./../root/./build/ir/Root.o
>>> defined in ./../root/./build/ir/Root.o

lake-duplicate-symbol.tar.gz

created time in 12 days

issue commentleanprover/lake

Breaks when changing git repo url

It would also be great if lake could show the actual error message from git instead of just the error code.

Seconded -- I get git exited with code 128 fairly often for seemingly different reasons, and it would be much easier to debug with a more detailed message.

gebner

comment created time in 12 days

push eventdselsam/lean3port

Daniel Selsam

commit sha 34fd0d560807009e3619a0193e20f86155287285

fix: update mathlib4 checksum

view details

push time in 13 days

push eventdselsam/lean3port

Daniel Selsam

commit sha 3b82cb90f1a79d825f794e093235ad1fbb4ce7e7

bump: nightly-2022-01-01

view details

push time in 13 days

fork dselsam/lean3port

Stub for downloading mathport artifacts for Lean 3

fork in 13 days

push eventdselsam/dselsam.github.io

Daniel Selsam

commit sha 4e9c94ec3eb6afcf54a0eccc9165766815ed0954

rm msft affiliation/contact

view details

push time in a month

issue commentleanprover-community/mathport

remove unnecessary implicit arguments in lambdas

Synport won't have type information in general. It might be easier to set_option hackport.disableImplicitLambda true on the Lean4 side and then (semi-auto) refactor once things type-check in Lean4.

semorrison

comment created time in a month

pull request commentleanprover-community/mathport

feat(CI): continuous integration to create nightly-* and pr-* releases

https://github.com/leanprover-community/mathport/commit/a78f4d4f7732504cb25245897fe1c352a3898a37

semorrison

comment created time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 069873c66eb6085ad6d5efcdc0158142726bd1f6

feat: tweaks to roundtrip with new CI releases

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha aea44e5435ad326e4223b33904dd79a0106abceb

wip: try roundtrip with new CI releases

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha afcd81135ab88969390d405081ca811b2e125d2d

wip: try roundtrip with new CI releases

view details

push time in 2 months

push eventleanprover-community/mathport

Scott Morrison

commit sha a78f4d4f7732504cb25245897fe1c352a3898a37

feat: initial attempt at CI

view details

Daniel Selsam

commit sha 7ea701f6272314399e51380ab72d86444a58a35e

feat: add lean-toolchain to Lean4Packages

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 4e083fe7c3f687ea13310eef8cfd59250182d212

chore: make importMathbin use mathbin

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 1803e075f38f530ccd957a11bac7b1ce49d0674f

fix: misc names/paths/links

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 9c397a5ccca0ad3ebaa4a3f24dee90c2119f1458

fix: misc names/paths/links

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 8dbcb284a01d70dc0de4dc91ef299a429fc6cd6e

fix: misc names/paths/links

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha e3522703dcd7f0f8070cfe8a8614113fff47de3a

fix: misc names/paths/links

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha 596b2623958d99f7621e4bb0afa95e7843ef153c

fix: standardize paths to ported lean4 oleans

view details

push time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha eb8c61c56137cad03cd64c45fdb6f585d02c21a7

fix: standardize paths to ported lean4 oleans

view details

push time in 2 months

CommitCommentEvent

push eventdselsam/lean4

Daniel Selsam

commit sha a9634f64dcc514c5c57e06576148f00b24581663

refactor: separate out Char.isAscii

view details

push time in 2 months

push eventdselsam/lean4

Daniel Selsam

commit sha 3c1c947602d989eac444fb8fe32ea1f111b2a627

feat: flag to minimize unicode

view details

push time in 2 months

create barnchdselsam/lean4

branch : minimizeUnicode

created branch time in 2 months

push eventleanprover-community/mathport

Daniel Selsam

commit sha ad83ed70e9e0480d082dce6c0cd0d6a500656cc6

refactor: rm Lib4, new dir for lean4 packages

view details

push time in 2 months

push eventdselsam/mathport

Daniel Selsam

commit sha ad83ed70e9e0480d082dce6c0cd0d6a500656cc6

refactor: rm Lib4, new dir for lean4 packages

view details

push time in 2 months

push eventdselsam/mathport

Daniel Selsam

commit sha d625b85a35e466e9a030fece3361edc0b96fd51a

refactor: rm Lib4, new dir for lean4 packages

view details

push time in 2 months

push eventdselsam/mathport

Daniel Selsam

commit sha 2a4881c5fbc429810874ef762ce2c9f9e4118c80

chore: finally rm mathport dependency

view details

Daniel Selsam

commit sha 8c70a1d50cd97d11fa142cb5bfd9b04b869282d2

refactor: rm Lib4, new dir for lean4 packages

view details

push time in 2 months

more