score_sde_probability_flow
Build deterministic probability-flow ODE diagnostics and one-dimensional likelihood corrections for Score SDE recovery.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- score_sde_probability_flow
- description
- Build deterministic probability-flow ODE diagnostics and one-dimensional likelihood corrections for Score SDE recovery.
02 资源文件 2
- 脚本scripts/probability_flow.py1.6 KB
- 脚本tests/test_probability_flow.py606 B
03 SKILL.md 正文
Score SDE Probability Flow
Use this skill when a recovery experiment must show the deterministic ODE counterpart of the reverse SDE or a likelihood-style divergence check.
Inputs
- State value.
- Time value.
- Forward drift and diffusion or simple VP-like coefficients.
- Score and score derivative for exact low-dimensional tests.
Outputs
- ODE drift.
- Zero diffusion flag.
- Divergence and log-density correction.
- Finite-step diagnostic.
Workflow
- Compute probability-flow drift with half the reverse-SDE score correction.
- Set diffusion to zero.
- Compute exact one-dimensional divergence when using a linear score.
- Apply one Euler step and accumulate the negative divergence as a log-density correction.
- Record that this is a proxy diagnostic unless full image scaling is provided.
Validation
Run:
python tests/test_probability_flow.py
python scripts/probability_flow.py --x 1.0 --t 0.5 --score-coeff -1.0
Limitations
This skill validates probability-flow mechanics; it does not report CIFAR-10 bits/dim without the full trained model and data pipeline.
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).