schrodinger_bridge_coupling
Build small entropy-regularized coupling checks and bridge variance schedules for SB conditional flow matching.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- schrodinger_bridge_coupling
- description
- Build small entropy-regularized coupling checks and bridge variance schedules for SB conditional flow matching.
02 资源文件 2
- 脚本scripts/sb_coupling.py1.4 KB
- 脚本tests/test_sb_coupling.py511 B
03 SKILL.md 正文
Schrödinger Bridge Coupling
Use this skill when a recovery or implementation needs a compact SB-CFM mechanism check: entropy-regularized pair weights and bridge variance behavior. Do not treat the simple row-normalized Gibbs helper as a full Sinkhorn solver for production transport.
Inputs
- Source and target minibatches.
- Positive entropy coefficient
epsilon. - Time
tand bandwidthsigmafor the bridge variance schedule.
Outputs
- Stabilized Gibbs weights derived from squared costs.
- Row-normalized coupling proxy for small tests.
- Bridge standard deviation
sigma * sqrt(t * (1 - t)).
Workflow
- Compute squared costs between source and target points.
- Convert costs to stabilized
exp(-cost / epsilon)weights. - Normalize rows for a lightweight concentration diagnostic.
- Compute bridge variance and verify endpoint behavior.
Validation
Run python tests/test_sb_coupling.py or the Distiller skill-tree validator with tests enabled.
Limitations
This skill supports reduced recovery checks. A full SB-CFM implementation should use Sinkhorn/IPF normalization and stochastic bridge sampling.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).