prompted_multitask_training
Execute a deterministic minimal optimizer step on prompted multitask text examples and log loss/parameter changes.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- prompted_multitask_training
- description
- Execute a deterministic minimal optimizer step on prompted multitask text examples and log loss/parameter changes.
02 资源文件 2
- 脚本scripts/tiny_train.py1.6 KB
- 脚本tests/test_tiny_train.py549 B
03 SKILL.md 正文
Prompted Multitask Training Step
Use this skill for bounded recovery of the T0/P3 mechanism when full encoder-decoder fine-tuning is not feasible. It provides a small softmax classifier over bag-of-words features so an experiment can prove that prompted examples are consumed by a real parameter update.
Inputs
- Rendered training records with
sourcetext andtargetlabels. - Label vocabulary and optional learning rate/epoch count.
Outputs
- Parameters before and after training, loss before/after, and a prediction helper.
Workflow
- Build a deterministic vocabulary from source-task prompted text.
- Initialize label weights to zero for reproducibility.
- Compute multiclass logistic loss and gradients.
- Apply gradient descent updates and log parameter changes.
Validation
Run the included tests. Recovery should save training_trace.json with params_before and params_after.
Limitations
This is a declared proxy for supervised prompted fine-tuning, not a replacement for T5/T0 model training.
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).