conditional_flow_matching_objective
Build conditional flow matching interpolation states, target velocities, and squared regression losses for bounded CNF recovery experiments.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- conditional_flow_matching_objective
- description
- Build conditional flow matching interpolation states, target velocities, and squared regression losses for bounded CNF recovery experiments.
02 资源文件 2
- 脚本scripts/cfm_objective.py2.6 KB
- 脚本tests/test_cfm_objective.py458 B
03 SKILL.md 正文
Conditional Flow Matching Objective
Use this skill when a recovery experiment or implementation needs the simulation-free CFM training target from the paper. Do not use it as an ODE sampler or as a replacement for evaluating a trained continuous normalizing flow.
Inputs
- Matched source and target point arrays
x0andx1with the same shape. - One scalar time or one time per sample in
[0, 1]. - A candidate vector-field prediction array with the same shape when computing loss.
Outputs
- Interpolated states
x_t = (1 - t) x0 + t x1. - Target velocities
u_t = x1 - x0for the deterministic linear path. - Mean squared velocity-regression loss and per-sample losses.
Workflow
- Validate that source and target arrays have identical batch and feature dimensions.
- Broadcast scalar time values or validate one time per sample.
- Construct linear conditional path samples and displacement targets.
- Compare predicted velocities against targets with squared error.
- Log whether smoothing/noise was omitted when using a reduced deterministic proxy.
Validation
Run python tests/test_cfm_objective.py from this skill directory, or validate the tree with the Distiller validate_skill_tree.py --run-tests command.
Limitations
This skill implements the deterministic mean-path target used by I-CFM and OT-CFM proxy experiments. It does not perform neural ODE integration, density evaluation, or stochastic Gaussian path 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).