feat: prove the whole trainer bit-exact in C (== model == Verilog) - #1878
Merged
Conversation
Extends cross-target bit-exactness from the GF-T primitives to the WHOLE trainer.
verify_trainer_c.py emits the microsequencer as a C program (C GF-T primitives via
t27c gen-c + a microcode interpreter + the operand-modifier modf), runs a full
80-step training run (forward+backprop+update), and checks every output per step
against the independent Python GF-T model: C trainer == model bit-exact for 2-layer
(2,2,1), multi-output (2,4,2), and deep [2,4,3,1]. Verilog == model is already
proven, so the entire training loop is bit-identical across {C, Verilog, model}.
Verified the guard catches a corrupted modf (9/80 mismatches). Wired into
emit-bitexact-gate (SKIPs if cc/t27c absent).
Refs #1764
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Extends cross-target bit-exactness from the GF-T primitives to the whole trainer.
verify_trainer_c.pyemits the microsequencer as a C program (C GF-T primitives via t27cgen-c+ a microcode interpreter + the operand-modifiermodf), runs a full 80-step training run (forward+backprop+update), and checks every output per step against the independent Python GF-T model: C trainer == model bit-exact for 2-layer (2,2,1), multi-output (2,4,2), and deep [2,4,3,1]. Verilog == model is already proven, so the entire training loop is bit-identical across {C, Verilog, model}. Verified the guard catches a corrupted modf (9/80 mismatches). Wired into emit-bitexact-gate. Refs #1764