Pinned Repositories
extract_prompt
prompt reverse engineering for anyone, completely locally, on a cpu
GPICL
A replication of General-Purpose In-Context Learning by Meta-Learning Transformers https://arxiv.org/abs/2212.04458
kir-gadjello.github.io
my personal blog
llm
a cli tool to make local & remote LLMs useful in the shell (bonus: streaming & interactivity supported)
mpa-eval
A very quick benchmark sensitive to quantization of advanced LLMs
picoagent-rnd
Web & CLI capable LLM agent (research prototype, no framework dependencies).
safer_unpickle
zipslicer
A library for incremental loading of large PyTorch checkpoints
kir-gadjello's Repositories
kir-gadjello/zipslicer
A library for incremental loading of large PyTorch checkpoints
kir-gadjello/picoagent-rnd
Web & CLI capable LLM agent (research prototype, no framework dependencies).
kir-gadjello/extract_prompt
prompt reverse engineering for anyone, completely locally, on a cpu
kir-gadjello/safer_unpickle
kir-gadjello/GPICL
A replication of General-Purpose In-Context Learning by Meta-Learning Transformers https://arxiv.org/abs/2212.04458
kir-gadjello/kir-gadjello.github.io
my personal blog
kir-gadjello/llm
a cli tool to make local & remote LLMs useful in the shell (bonus: streaming & interactivity supported)
kir-gadjello/mpa-eval
A very quick benchmark sensitive to quantization of advanced LLMs
kir-gadjello/ChatGPT-Paper-Reader
This repo offers a simple interface that helps you to read&summerize research papers in pdf format. You can ask some questions after reading. This interface is developed based on openai API and using GPT-3.5-turbo model.
kir-gadjello/clipgrep
A semantic image search tool bringing venerable grep into a new AI era. Search by Prompt!
kir-gadjello/infinity
Infinity is a high-throughput, low-latency REST API for serving vector embeddings, supporting a wide range of text-embedding models and frameworks.
kir-gadjello/llama.cpp-oaicompat
Port of Facebook's LLaMA model in C/C++
kir-gadjello/MemGPT
Building persistent LLM agents with long-term memory 📚🦙
kir-gadjello/notes
Note-taking application, write down your thoughts.
kir-gadjello/palace-of-memories
personal memory spatial association environment
kir-gadjello/HandheldHelper
A personal fully local LLM client with builtin inference engine. Desktop & mobile support. GPLv3-licensed.
kir-gadjello/HungarianMathEval
An implementation of fully-automatic eval & grading for Hungarian high school math eval for LLMs
kir-gadjello/llamacpp-embed
A customized llama.cpp inference runtime I use for my desktop and mobile apps
kir-gadjello/llm_fns
Ultra-minimalistic unobtrusive functional LLM API abstractions
kir-gadjello/MiniCPM-V
MiniCPM-Llama3-V 2.5: A GPT-4V Level MLLM on Your Phone
kir-gadjello/obsidian-omnisearch
A personal fork of omniserch, a search engine that "just works" for Obsidian.
kir-gadjello/reflex-chat
A ChatGPT clone built in Reflex
kir-gadjello/RepoDreamer
A synthetic data generation engine for teaching LLMs project-level software engineering skills
kir-gadjello/SillyTavern
LLM Frontend for Power Users.
kir-gadjello/tabbyAPI
An OAI compatible exllamav2 API that's both lightweight and fast
kir-gadjello/text-embeddings-inference
A blazing fast inference solution for text embeddings models
kir-gadjello/vlcn-docs
vlcd docs site
kir-gadjello/WorkBench
WorkBench: a Benchmark Dataset for Agents in a Realistic Workplace Setting.