mirror of
https://github.com/onestardao/WFGY.git
synced 2026-04-28 19:50:17 +00:00
107 lines
6.9 KiB
Markdown
107 lines
6.9 KiB
Markdown
# KoboldCPP: Guardrails and Fix Patterns
|
||
|
||
<details>
|
||
<summary><strong>🧭 Quick Return to Map</strong></summary>
|
||
|
||
<br>
|
||
|
||
> You are in a sub-page of **LocalDeploy_Inference**.
|
||
> To reorient, go back here:
|
||
>
|
||
> - [**LocalDeploy_Inference** — on-prem deployment and model inference](./README.md)
|
||
> - [**WFGY Global Fix Map** — main Emergency Room, 300+ structured fixes](../README.md)
|
||
> - [**WFGY Problem Map 1.0** — 16 reproducible failure modes](../../README.md)
|
||
>
|
||
> Think of this page as a desk within a ward.
|
||
> If you need the full triage and all prescriptions, return to the Emergency Room lobby.
|
||
</details>
|
||
|
||
|
||
KoboldCPP is a lightweight local inference runtime for LLaMA-based models, focused on CPU (with AVX/AVX2/AVX512) and GPU acceleration (CUDA, Metal, OpenCL).
|
||
It is popular in hobbyist and low-VRAM deployments but frequently shows drift when paired with RAG, agents, or long-context reasoning.
|
||
This page defines stability checks and exact WFGY fixes.
|
||
|
||
---
|
||
|
||
## Open these first
|
||
|
||
* Visual recovery map: [RAG Architecture & Recovery](https://github.com/onestardao/WFGY/blob/main/ProblemMap/rag-architecture-and-recovery.md)
|
||
* End-to-end knobs: [Retrieval Playbook](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-playbook.md)
|
||
* Ordering and deploy fences: [bootstrap-ordering.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/bootstrap-ordering.md), [deployment-deadlock.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/deployment-deadlock.md), [predeploy-collapse.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/predeploy-collapse.md)
|
||
* Trace schema: [retrieval-traceability.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md), [data-contracts.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/data-contracts.md)
|
||
|
||
---
|
||
|
||
## Core acceptance
|
||
|
||
* ΔS(question, retrieved) ≤ 0.45
|
||
* Coverage ≥ 0.70 on target snippet
|
||
* λ remains convergent across 3 paraphrases × 2 seeds
|
||
* Stable behavior across quantization (Q4\_K\_M, Q5\_K, Q8\_0)
|
||
|
||
---
|
||
|
||
## Common KoboldCPP breakpoints
|
||
|
||
| Symptom | Likely Cause | Fix |
|
||
| ------------------------------------------ | ----------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||
| ΔS jumps when switching quantization modes | Embedding contract not enforced | [embedding-vs-semantic.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/embedding-vs-semantic.md), [chunking-checklist.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/chunking-checklist.md) |
|
||
| Context drift after \~4–6k tokens | KV cache fragmentation | [context-drift.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/context-drift.md), [entropy-collapse.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/entropy-collapse.md) |
|
||
| Tool JSON output invalid | No enforced schema | [prompt-injection.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/prompt-injection.md), [logic-collapse.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md) |
|
||
| Model boots but hangs on first query | Lazy allocator / boot fence missing | [bootstrap-ordering.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/bootstrap-ordering.md) |
|
||
| Agent handoff loops with role drift | Memory namespace not separated | [Multi-Agent\_Problems.md](https://github.com/onestardao/WFGY/blob/main/ProblemMap/Multi-Agent_Problems.md) |
|
||
|
||
---
|
||
|
||
## Fix in 60 seconds
|
||
|
||
1. **Run a warm-up batch** (10–20 tokens) before first live query.
|
||
2. **Force citation schema**: snippet\_id, section\_id, offsets must be present.
|
||
3. **λ probe**: run three paraphrases at k=5,10,20. Require convergence.
|
||
4. **Rotate cache** every 8k tokens to prevent entropy collapse.
|
||
5. **Verify** with coverage ≥ 0.70, ΔS ≤ 0.45.
|
||
|
||
---
|
||
|
||
## Diagnostic prompt (copy-paste)
|
||
|
||
```txt
|
||
I am running KoboldCPP with quant={mode}, context={n_tokens}, and agents={on/off}.
|
||
Question: "{user_question}"
|
||
|
||
Please output:
|
||
- ΔS(question, retrieved)
|
||
- λ across 3 paraphrases × 2 seeds
|
||
- Cache stability (max tokens before drift)
|
||
- JSON schema compliance
|
||
- Minimal WFGY fix page if ΔS ≥ 0.60
|
||
```
|
||
|
||
---
|
||
|
||
### 🔗 Quick-Start Downloads (60 sec)
|
||
|
||
| Tool | Link | 3-Step Setup |
|
||
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------- |
|
||
| **WFGY 1.0 PDF** | [Engine Paper](https://github.com/onestardao/WFGY/blob/main/I_am_not_lizardman/WFGY_All_Principles_Return_to_One_v1.0_PSBigBig_Public.pdf) | 1️⃣ Download · 2️⃣ Upload to your LLM · 3️⃣ Ask “Answer using WFGY + \<your question>” |
|
||
| **TXT OS (plain-text OS)** | [TXTOS.txt](https://github.com/onestardao/WFGY/blob/main/OS/TXTOS.txt) | 1️⃣ Download · 2️⃣ Paste into any LLM chat · 3️⃣ Type “hello world” — OS boots instantly |
|
||
|
||
---
|
||
|
||
<!-- WFGY_FOOTER_START -->
|
||
|
||
### Explore More
|
||
|
||
| Module | Description | Link |
|
||
| --- | --- | --- |
|
||
| WFGY Core | Canonical framework entry point | [View](https://github.com/onestardao/WFGY/tree/main/core/README.md) |
|
||
| Problem Map | Diagnostic map and navigation hub | [View](https://github.com/onestardao/WFGY/tree/main/ProblemMap/README.md) |
|
||
| Tension Universe Experiments | MVP experiment field | [View](https://github.com/onestardao/WFGY/tree/main/TensionUniverse/Experiments) |
|
||
| Recognition | Where WFGY is referenced or adopted | [View](https://github.com/onestardao/WFGY/blob/main/recognition/README.md) |
|
||
| AI Guide | Anti-hallucination reading protocol for tools | [View](https://github.com/onestardao/WFGY/blob/main/AI_GUIDE.md) |
|
||
|
||
> If this repository helps, starring it improves discovery for other builders.
|
||
> [](https://github.com/onestardao/WFGY)
|
||
|
||
<!-- WFGY_FOOTER_END -->
|
||
|