settings-and-configuration
Configure PaperQA Settings, named configs, LLM providers, embeddings, vector stores, prompts, callbacks, and optional dependencies.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- settings-and-configuration
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Configure PaperQA Settings, named configs, LLM providers, embeddings, vector stores, prompts, callbacks, and optional dependencies.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/model-and-embedding-config.md11.1 KB
- 参考文档references/settings-reference.md12.4 KB
- 参考文档references/troubleshooting.md6.4 KB
- 脚本scripts/print_named_settings.py6.8 KB
- 脚本scripts/validate_settings_json.py11.6 KB
03 SKILL.md 正文
PaperQA Settings and Configuration
Use this sub-skill when the task is to choose, validate, or troubleshoot PaperQA configuration rather than run a query or build an index.
Route here for
- Loading or adapting named
Settingsconfigs such asfast,high_quality,debug, clinical-trials configs,contracrow,wikicrow,openreview, or OpenAI tier-limit configs. - Setting all PaperQA model roles:
llm,summary_llm,agent.agent_llm, and parser enrichment LLMs. - Choosing embedding providers, hybrid/sparse/local embeddings, vector stores, embedding cache behavior, and Qdrant prerequisites.
- Writing or checking PaperQA settings JSON, nested settings dictionaries, prompt templates, LiteLLM router configs, rate limits, and callback-related settings.
- Diagnosing API-key/provider mismatches, optional extras, prompt variable validation, and
gpt-5/o1temperature behavior.
Route elsewhere
- To actually ask questions, call
Docs.aquery, gather evidence, or use callbacks during a run, use../agentic-rag/. - For
pqacommands, manifest/index operation, and search-index lifecycle, use../cli-and-indexing/. - For parser function selection, PDF/media parsing, Office parsing, and chunking details, use
../docs-and-parsing/. - For ClinicalTrials.gov, OpenReview, Zotero, and metadata-provider semantics, use
../metadata-and-sources/.
Operating workflow
- Start from
Settings()orSettings.from_name("<config>"); use settings-reference.md to decide which fields belong in root,answer,parsing,prompts, oragent. - If changing providers, update every active model role together:
llm,summary_llm,agent.agent_llm,embedding, and parserenrichment_llmwhen multimodal enrichment is on. Use model-and-embedding-config.md. - Validate settings JSON before a run with
scripts/validate_settings_json.py; print bundled config summaries withscripts/print_named_settings.py. - If validation or provider setup fails, use troubleshooting.md before changing query code.
Bundled safe helpers
scripts/print_named_settings.pyprints concise summaries of installed bundled settings without calling LLMs, embeddings, network services, or external automation.scripts/validate_settings_json.pyvalidates a PaperQA settings JSON file and reports provider/key, prompt-variable, router-shape, temperature, and optional-extra warnings without calling LLMs or embeddings.
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).