README
FPRM — Fixed-Point Tiny Recursive Models
Checkpoints for FPRM / FPTRM (Fixed-Point Tiny Recursive Model), a weight-tied iterative reasoner trained with a fixed-point solver and test-time compute scaling.
Each subfolder is one self-contained run (checkpoints + exact config + model source +
reproduction scripts). See each folder's README.md for the full recipe and results.
| folder | task | best test metric | notes |
|---|---|---|---|
maze/ | Maze-Hard 30×30 | 87.0% exact-match | single-z, conv1d/k4, non-augmented; eval @ max_iter 35k, decay 0.996/pat 10 |
sudoku/ | Sudoku-Extreme | 94.2% exact-match | single-z, conv2d/k3, norm-placement=none; eval @ max_iter 35k, decay 0.997/pat 10 |
arc1/ | ARC-1-concept (aug-1000) | 47.5% pass@2 | single-z, conv1d/k4, norm-placement=none; eval @ max_iter 1000 |
arc2/ | ARC-2-concept (aug-1000) | 6.2% pass@2 | single-z, conv1d/k4, norm-placement=none; eval @ max_iter 1000 |
Checkpoints are saved every 5000 epochs as step_<N> (EMA-averaged eval weights, the ones
to load) and step_<N>_train_state.pt (full training state for resuming). The arc1/ and
arc2/ folders ship eval checkpoints only (no train_state).
Contributors
8 commits
8 commits
Similar repos
README
FPRM — Fixed-Point Tiny Recursive Models
Checkpoints for FPRM / FPTRM (Fixed-Point Tiny Recursive Model), a weight-tied iterative reasoner trained with a fixed-point solver and test-time compute scaling.
Each subfolder is one self-contained run (checkpoints + exact config + model source +
reproduction scripts). See each folder's README.md for the full recipe and results.
| folder | task | best test metric | notes |
|---|---|---|---|
maze/ | Maze-Hard 30×30 | 87.0% exact-match | single-z, conv1d/k4, non-augmented; eval @ max_iter 35k, decay 0.996/pat 10 |
sudoku/ | Sudoku-Extreme | 94.2% exact-match | single-z, conv2d/k3, norm-placement=none; eval @ max_iter 35k, decay 0.997/pat 10 |
arc1/ | ARC-1-concept (aug-1000) | 47.5% pass@2 | single-z, conv1d/k4, norm-placement=none; eval @ max_iter 1000 |
arc2/ | ARC-2-concept (aug-1000) | 6.2% pass@2 | single-z, conv1d/k4, norm-placement=none; eval @ max_iter 1000 |
Checkpoints are saved every 5000 epochs as step_<N> (EMA-averaged eval weights, the ones
to load) and step_<N>_train_state.pt (full training state for resuming). The arc1/ and
arc2/ folders ship eval checkpoints only (no train_state).
Similar repos
Linked in READMEs
Contributors
8 commits
8 commits