forward_diffusion_schedule
Build DDPM beta schedules and closed-form forward noising coefficients for 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
- forward_diffusion_schedule
- description
- Build DDPM beta schedules and closed-form forward noising coefficients for recovery experiments.
02 资源文件 3
03 SKILL.md 正文
forward_diffusion_schedule
Use this skill when a recovery or implementation needs the DDPM mechanism from Ho et al. rather than a generic denoising routine.
Inputs
Provide explicit numeric arrays or scalar fixtures, beta schedule parameters, timestep ids, and any predictor parameters required by the module. Inputs must not come from the original source repository during recovery.
Outputs
Return deterministic JSON-serializable numeric evidence: coefficients, sampled noisy values, losses, posterior statistics, or recovery traces depending on the module.
Workflow
- Validate schedule length, beta range, shapes, and timestep bounds.
- Execute the module's DDPM equation directly with deterministic fixtures when possible.
- Record enough intermediate values for downstream recovery checks.
- Keep full-result claims separate from reduced/proxy evidence.
Validation
Run python -m pytest tests or python scripts/ddpm_utils.py from this skill directory. The bundled tests use only the Python standard library.
Limitations
This skill captures the transferable paper mechanism and small deterministic tests. It does not include the original TensorFlow model, cloud data pipeline, or pretrained CIFAR-10 checkpoints.
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).