TGTGInsighttelegram intelligenceLIVE / telegram public index
← Python Заметки

TGINSIGHT SIMILAR POSTS

Најди сличен содржај

Изворен канал @pythonotes · Post #425 · 20 апр.

Недавно делал быстрый прототип асинхронного приложения в котором требовалось вызывать много синхронного кода. Да, я знаю, что это не лучший дизайн, но нужно было быстрое решение на один процесс и без очередей. Поэтому я выполнял код в потоках. Выглядело это примерно так: from fastapi.concurrency import run_in_threadpool async def execute(data: DataRequest) -> DataResponse: try: result = await run_in_threadpool(sync_function, data) return DataResponse(data=result) except Exception as e: return DataResponse( error=str(e), success=False, ) В общем работает нормально. Для всех вызовов под капотом используется общий тредпул, всё работает предсказуемо. Но потребовалось изменить количество запускаемых в пуле потоков (по умолчанию создается 40 воркеров). Так как дело происходит с FastAPI, делается это через lifespan используя настройки anyio: import anyio @asynccontextmanager async def lifespan(app: FastAPI): limiter = anyio.to_thread.current_default_thread_limiter() limiter.total_tokens = 100 yield # если вдруг нужно вернуть обратно limiter.total_tokens = 40 Зачем менять количество воркеров? - уменьшить, если оперативки мало (один тред занимает ~8мб) - увеличить чтобы выдержать нагрузку Если есть предложения получше при тех же вводных - предлагайте😉 #async

Hashtags

Резултати

Пронајдени 2 слични објави

Пребарај: #transparencyinai

当前筛选 #transparencyinai清除筛选
AI & Law

@ai_and_law · Post #417 · 11.10.2024 г., 07:04

AI Copyright Lawsuit: A New Era of Dataset Transparency? In a groundbreaking move for AI copyright disputes, a judge has permitted authors' representatives to inspect OpenAI's confidential training dataset under strict conditions. This decision could set a precedent for how training data is scrutinized in future legal battles over AI-generated content. The inspection is to be conducted in a highly controlled environment—either at OpenAI’s San Francisco office or a secure location within a 25-mile radius. A secured, internet-free computer will be provided, containing a README file with a directory of the data. The inspection room bans recording devices, although a note-taking computer can be provided under supervision, ensuring that sensitive data remains protected. Only individuals qualified under the protective order can access the data, emphasizing the heightened confidentiality. This level of transparency and rigor in examining training data might become a standard in AI litigation. As AI technology continues to reshape industries, balancing innovation with intellectual property rights will require similar measures to ensure fairness and accountability. #AICopyright#TransparencyInAI#AIRegulation

AI & Law

@ai_and_law · Post #197 · 22.12.2023 г., 08:04

CAI Unveils Public "Zero Draft" for AI Framework Convention Greetings AI & Law community! The European Committee on Artificial Intelligence (CAI) has announced the public release of the revised "Zero Draft" Framework Convention on Artificial Intelligence, Human Rights, Democracy, and the Rule of Law. Key Points: 1️⃣ The "Zero Draft" serves as the foundational document for drafting the Framework Convention and is not the final outcome of CAI negotiations. 2️⃣ A Consolidated Working Draft has been prepared based on the first reading of the revised Zero Draft, containing preliminary agreements and proposals for further negotiations. 3️⃣ The text which is not in brackets and in red font is considered to be preliminarily agreed. 4️⃣ The CAI Bureau has decided to make this document public, providing transparency into the ongoing process. The recently released Draft Framework Convention, based on outcomes from the second reading, will be the focal point for the upcoming 9th Plenary meeting. Notably, provisions under negotiation include the Scope, Risk and impact management framework, and the Conference of the Parties. #AIlaw#CAI#AIregulation#FrameworkConvention#TransparencyInAI