diff options
| author | ericmarin <maarin.eric@gmail.com> | 2026-03-21 11:47:40 +0100 |
|---|---|---|
| committer | ericmarin <maarin.eric@gmail.com> | 2026-03-21 12:00:16 +0100 |
| commit | e2abe9d9ec649b849cc39b516c1db1b4fa592003 (patch) | |
| tree | d74dcc2e0691bb587d2a9a695639517d3aec9256 /notes.norg | |
| parent | af4335cf47984576e7493a0eb6569d3f6ecc31c8 (diff) | |
| download | vein-e2abe9d9ec649b849cc39b516c1db1b4fa592003.tar.gz vein-e2abe9d9ec649b849cc39b516c1db1b4fa592003.zip | |
created class
Diffstat (limited to 'notes.norg')
| -rw-r--r-- | notes.norg | 6 |
1 files changed, 3 insertions, 3 deletions
@@ -4,14 +4,14 @@ description: WIP tool to prove NNEQ using Interaction Nets as pre-processor fo m authors: ericmarin categories: research created: 2026-03-14T09:21:24 -updated: 2026-03-18T17:46:56 +updated: 2026-03-21T11:53:27 version: 1.1.1 @end * TODO - - (!) Scalability: %Impossible with Inpla, I would need to implement my own engine% + - (?) Scalability: Added cache for already processed NN. Inpla is the bottleneck but now I removed most of the limits. - (x) Soundness of translated NN: {:proof.norg:}[PROOF] - - ( ) Compatibility with other types of NN + - (x) Compatibility with other types of NN: Added support for ONNX format. For now only GEMM and ReLU operations are implemented. - ( ) Comparison with other tool ({https://github.com/NeuralNetworkVerification/Marabou}[Marabou], {https://github.com/guykatzz/ReluplexCav2017}[Reluplex]) - ( ) Add Range agent to enable ReLU optimization |