Минш: почти линейное время глобального выравнивания строк в нескольких линиях Python
MinSH: Near-linear time global string alignment in a few lines of Python .
GitHub
GitHub - pesho-ivanov/minSH: A minimal Python re-implementation of the A* with seed heuristic for exact global alignmnet (edit…
A minimal Python re-implementation of the A* with seed heuristic for exact global alignmnet (edit distance) in near-linear time - pesho-ivanov/minSH
Показать HN: Определите и реализуйте любую функцию на лету с LLMS
Show HN: Define and implement any function on the fly with LLMs .
GitHub
GitHub - pycui/universal-api: Define and implement any functions on the fly with LLMs
Define and implement any functions on the fly with LLMs - pycui/universal-api
Show HN: GPTlint-стандарт OSS для обзора кода на основе LLM
Show HN: GPTLint – an OSS standard for LLM-based code review .
GitHub
GitHub - gptlint/gptlint: A linter with superpowers! 🔥 Use LLMs to enforce best practices across your codebase.
A linter with superpowers! 🔥 Use LLMs to enforce best practices across your codebase. - gptlint/gptlint
DOCLITE: PHP NOSQL База данных и хранилище документов с использованием SQLite
Doclite: PHP NoSQL database and document store using SQLite .
GitHub
GitHub - dwgebler/doclite: PHP NoSQL database and document store
PHP NoSQL database and document store. Contribute to dwgebler/doclite development by creating an account on GitHub.
Показать HN: Thru - Создать туннели Quic/TCP/UDP друг для друга
Show HN: Thru – Create QUIC/TCP/UDP tunnels for one another .
GitHub
GitHub - plestoon/thru: A QUIC tunnel for TCP/UDP
A QUIC tunnel for TCP/UDP. Contribute to plestoon/thru development by creating an account on GitHub.
Показать HN: найти аналогичные папки на основе имени папки, размера папки и COUNT
Show HN: Find similar folders based on folder name, folder size, and count .
GitHub
GitHub - chapmanjacobd/library: 99+ CLI tools to build, browse, and blend your media library
99+ CLI tools to build, browse, and blend your media library - chapmanjacobd/library
Show HN: Я сделал новый клиент Lavalink с node.js
Show HN: I made a new LavaLink client with Node.js .
GitHub
GitHub - Fyphen1223/TsumiLink: Fast, easy-to-use, reliable, feature-rich Lava/NodeLink compatible client.
Fast, easy-to-use, reliable, feature-rich Lava/NodeLink compatible client. - Fyphen1223/TsumiLink
Принципы приглашения CHATGPT (поддержаны исследованиями)
ChatGPT Prompt Principles (backed by research) .
GitHub
GitHub - VILA-Lab/ATLAS: A principled instruction benchmark on formulating effective queries and prompts for large language models…
A principled instruction benchmark on formulating effective queries and prompts for large language models (LLMs). Our paper: https://arxiv.org/abs/2312.16171 - VILA-Lab/ATLAS
GGML Flash Внимание Поддержка, объединенная в llama.cpp
GGML Flash Attention support merged into llama.cpp .
GitHub
ggml : add Flash Attention by ggerganov · Pull Request #5021 · ggerganov/llama.cpp
ref #3365
Setting up what's needed for Flash Attention support in ggml and llama.cpp
The proposed operator performs:
// new
res = ggml_flash_attn(ctx, q, k, v, kq_mask, kq_scale);
// fused...
Setting up what's needed for Flash Attention support in ggml and llama.cpp
The proposed operator performs:
// new
res = ggml_flash_attn(ctx, q, k, v, kq_mask, kq_scale);
// fused...