feat(ilp): Tensorized Differentiable ILP engine - #1

Merged

levi770 merged 21 commits into

mainfrom

feat/tensorized-ilp

Feb 25, 2026

Commits on Feb 25, 2026

feat(ilp): add learnable_rule grammar production
levi770
committed
6df0165
feat(ilp): add LearnableRule AST type and Program field
levi770
committed
3e6f8e1
feat(ilp): parse learnable rules into LearnableRule AST
levi770
committed
2fe5fd5
feat(ilp): add learnable rule edges to dependency graph
levi770
committed
877ea64
feat(ilp): add TensorMaskedJoin RirNode variant
levi770
committed
f1ec5bc
feat(ilp): add TensorMaskedJoin arms in optimizer
levi770
committed
e274365
feat(ilp): lower LearnableRule to TensorMaskedJoin via lower_program
levi770
committed
fe9779e
feat(ilp): add IlpRegistry, executor fields, and accessors
levi770
committed
c4f3239
feat(ilp): add extract_nonzero_indices CUDA kernel and provider wrapper
levi770
committed
69ada5b
feat(ilp): implement execute_tensor_masked_join with hash-join dispatch
levi770
committed
15b9b2a
test(ilp): add CUDA kernel extraction tests
levi770
committed
0da335d
test(ilp): add end-to-end integration tests (T3.1-T3.3) with fact loading
levi770
committed
2464a37
feat(ilp): add CompiledIlpProgram PyO3 bindings for ILP
levi770
committed
da58694
test(ilp): add Python integration tests for ILP gradient flow
levi770
committed
fc24ec6
test(ilp): add additional RFC test coverage (M1-M5 gaps)
levi770
committed
3149a8a
fix(ilp): arity guard in TensorMaskedJoin + validation cleanup
levi770
committed
6b02bf8
docs(ilp): add tensorized ILP RFC and implementation plan
levi770
committed
d8f950a
fix(ilp): address PR review comments
levi770
committed
603d3a8
fix(ilp): keep simple torch.cuda skip gate for ILP tests
levi770
committed
7da2d6c
fix(ilp): per-fact credit projection, head validation, schema_size check
levi770
committed
deb6608
test(ilp): cover projected credit path, head validation, schema_size check
levi770
committed
72262fb

Navigation Menu

Uh oh!

feat(ilp): Tensorized Differentiable ILP engine - #1

feat(ilp): Tensorized Differentiable ILP engine#1
levi770 merged 21 commits into
mainfrom
feat/tensorized-ilp

Commits on Feb 25, 2026

feat(ilp): add learnable_rule grammar production

feat(ilp): add LearnableRule AST type and Program field

feat(ilp): parse learnable rules into LearnableRule AST

feat(ilp): add learnable rule edges to dependency graph

feat(ilp): add TensorMaskedJoin RirNode variant

feat(ilp): add TensorMaskedJoin arms in optimizer

feat(ilp): lower LearnableRule to TensorMaskedJoin via lower_program

feat(ilp): add IlpRegistry, executor fields, and accessors

feat(ilp): add extract_nonzero_indices CUDA kernel and provider wrapper

feat(ilp): implement execute_tensor_masked_join with hash-join dispatch

test(ilp): add CUDA kernel extraction tests

test(ilp): add end-to-end integration tests (T3.1-T3.3) with fact loading

feat(ilp): add CompiledIlpProgram PyO3 bindings for ILP

test(ilp): add Python integration tests for ILP gradient flow

test(ilp): add additional RFC test coverage (M1-M5 gaps)

fix(ilp): arity guard in TensorMaskedJoin + validation cleanup

docs(ilp): add tensorized ILP RFC and implementation plan

fix(ilp): address PR review comments

fix(ilp): keep simple torch.cuda skip gate for ILP tests

fix(ilp): per-fact credit projection, head validation, schema_size check

test(ilp): cover projected credit path, head validation, schema_size check