Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add HorizonRobotics/OE-Skills --skill j6-plugin-set-fake-quantizegit clone --depth 1 https://github.com/HorizonRobotics/OE-SkillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/horizonrobotics/oe-skills/j6-plugin-set-fake-quantize)<a href="https://agentmods.dev/skills/horizonrobotics/oe-skills/j6-plugin-set-fake-quantize"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/j6-plugin-set-fake-quantize/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/horizonrobotics/oe-skills/j6-plugin-set-fake-quantize"><img src="https://agentmods.dev/badge/skills/horizonrobotics/oe-skills/j6-plugin-set-fake-quantize.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00054 | $0.00833 |
| Opus 5.5 | $0.00022 | $0.00333 |
| Sonnet 5.5 | $0.00011 | $0.00167 |
| Haiku 4.5 | $0.00005 | $0.00083 |
Grade A, and why
j6-plugin-set-fake-quantize scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
为 Horizon 量化流程设置 fake quantize 状态(set_fake_quantize)
目标
在使用 horizon_plugin_pytorch 的 Calibration / QAT / Validation 流程时,按阶段为模型设置正确的 fake quantize 状态:
- Calibration 前:
CALIBRATION - QAT(训练)前:
QAT - Validation(评估)前:
VALIDATION
本 Skill 只做一件事:在合适位置添加/调用
horizon.quantization.set_fake_quantize(model, horizon.quantization.FakeQuantState.<STATE>)
不引入任何其他改动(不改模型结构、不改 qconfig、不改 prepare/convert、不改训练/数据逻辑)。
标准改法(通用模板)
1) 选择状态枚举值
fake quantize 有三种状态:
class FakeQuantState(Enum):
QAT = "qat"
CALIBRATION = "calibration"
VALIDATION = "validation"
语义与行为约束(按官方说明):
CALIBRATION:仅观测各算子输入/输出统计量(observer 统计)。QAT:观测统计量 + 执行伪量化(fake quant)。VALIDATION:仅执行伪量化,不再观测统计量。
2) 在阶段入口处调用 set_fake_quantize
Calibration 前
model.eval()
horizon.quantization.set_fake_quantize(
model, horizon.quantization.FakeQuantState.CALIBRATION
)
QAT(训练)前
horizon.quantization.set_fake_quantize(
model, horizon.quantization.FakeQuantState.QAT
)
Validation(评估)前
model.eval()
horizon.quantization.set_fake_quantize(
model, horizon.quantization.FakeQuantState.VALIDATION
)
关键注意事项(必须遵守)
1) Calibration 状态下不要再调用 model.eval()
一旦设置为 FakeQuantState.CALIBRATION,请勿再使用 model.eval(),否则将无法正常进行校准。
如果你的流程需要在校准时切换到 eval,请改为:
- 先
model.eval()(如果你确实需要 eval) - 再
set_fake_quantize(..., CALIBRATION)
并确保后续不再重复调用 model.eval()。
2) Validation 的顺序要求
Validation 推荐固定顺序:
- 先
model.eval() - 再
set_fake_quantize(..., VALIDATION)
以保证评估时不再更新统计量,仅执行伪量化。
3) 不要混用状态(尤其是训练/评估循环)
如果你的代码存在多个入口(如:训练脚本、校准脚本、导出脚本),每个入口都要在其对应阶段入口处显式设置状态,避免:
- QAT 训练时仍处于
CALIBRATION - Validation 评估时仍处于
QAT
快速自检清单
- Calibration 前调用:
set_fake_quantize(model, CALIBRATION),且之后不再model.eval()。 - QAT 训练前调用:
set_fake_quantize(model, QAT)。 - Validation 前调用:
model.eval()后紧接set_fake_quantize(model, VALIDATION)。
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 106 lines · 54 tokens per session scan A d0b93d749046
j6-plugin-set-fake-quantize is a skill published in the GitHub repository HorizonRobotics/OE-Skills (27 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 54 tokens to every session and 833 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
j6-plugin-set-fake-quantize compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| j6-plugin-set-fake-quantizeHorizonRobotics/OE-Skills | Skill | 27 | 4mo ago | Apache-2.0 | A | 54 | Claude Code, Codex |
| agent-platform-model-registrygoogle/skills | Skill | 22k | yesterday | Apache-2.0 | A | 60 | Claude Code, Codex |
| foundry-config-setupmicrosoft/agent-framework | Skill | 15k | yesterday | MIT | A | 65 | Claude Code, Codex |
| google-cloud-solution-agentic-analytics-spark-knowledge-cataloggoogle/skills | Skill | 22k | yesterday | Apache-2.0 | A | 132 | Claude Code, Codex |
| training-checkwanshuiyin/Auto-claude-code-research-in-sleep | Skill | 19k | 3d ago | MIT | A | 35 | Claude Code |
| nemo-automodel-launcher-configNVIDIA/skills | Skill | 3.9k | yesterday | Apache-2.0 | A | 30 | Claude Code |
| finetuningmicrosoft/skills | Skill | 3.2k | yesterday | MIT | A | 132 | Claude Code |