Lists (4)
Starred repositories
Beautifully designed wrapper around Tinker for browsing, probing, and switching LoRA checkpoints across experiments.
SWE-QA-Pro: A Representative Benchmark and Scalable Training Recipe for Repository-Level Code Understanding [ACL 2026]
Official repository of paper "Learning Project-wise Subsequent Code Edits via Interleaving Neural-based Induction and Tool-based Deduction"
🔄️ Sync Claude Code transcripts to SQLite for analytics, uses node:sqlite
A powerful MCP toolkit for coding, providing semantic retrieval and editing capabilities - the IDE for your agent
Dish: Deployment Imitating SWE-Agent Harness
Data and evaluation scripts for "CodePlan: Repository-level Coding using LLMs and Planning", FSE 2024
Manipulate Claude Code's 5-hour usage window into resetting when you actually need it.
Indexed regex search for large codebases, powered by trigram / sparse n‑gram indexes. A grep-like CLI that builds a local on-disk index, so searches stay fast even on huge monorepos.
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, …
James' cookbook of evaluations and finetuning experiments
We're back! Now firing notebooks out of a t-shirt gun.
Python library to add support for embedding natural code in Python with shared program state.
🎨 NeMo Data Designer: Generate high-quality synthetic data from scratch or from seed data.
Train LLMs and SLMs while you code with Claude or other Bash agents. BashGym enables you to collect the data from your coding sessions, create synthetic data, train LoRA adapters, and deploy open s…
Post-training with Tinker
Twinkle✨: Training workbench to make your model glow.
A JAX backend for Apple Metal Performance Shaders (MPS), enabling GPU-accelerated JAX computations on Apple Silicon.
CLI tool for running JAX training on Google Cloud Spot TPUs with automatic preemption recovery and checkpoint resumption