Repository navigation
spec(hslm): train_hslm counts 1948617 ternary trits in 19 matrices, with a test (Closes #10839) - #10840
Merged
Conversation
gHashTag
force-pushed
the
fix/hslm-train-ternary-count
branch
from
October 19, 2026 04:27
2b87b5b to
6cd4882
Compare
This was referenced Oct 19, 2026
Merged
Contributor
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #10839
The header of
specs/ml/train_hslm.t27said "19 matrices, 1771470 trits". 1771470 is only the 18 block matrices; with the 243 x 729 output head (177147) the 19 ternary matrices hold 1,948,617 trits, the numberspecs/hf/model_card_hslm.t27asserts as PARAMS_TERNARY (#10836).tr_ternary_count/tr_ternary_matricessum the spec's own segment table; new testthe_ternary_matrices_hold_1948617_tritsasserts 19 matrices, 1948617 trits, and the closed form 3 x (4 x 243 x 243 + 2 x 243 x 729) + 243 x 729.ml_train_hslmseal refreshed.Measured on the Railway fs-lab, not the Mac: test-report on
specs/ml/train_hslm.t27is 21/21 pass, vacuous passes 0 of 21; seal --save run there, and the seal file was checked by sha256 at both ends.specs/ml/eval_hslm.t27is not on master yet, so there is no second seal to refresh; it must be resealed when it lands on top of this change.🤖 Generated with Claude Code