CS @ UC Berkeley | @pytorch
Repositories
theap06/tensordict
TensorDict is a pytorch dedicated tensor container.
theap06/tinker-cookbook
Post-training with Tinker
theap06/IsaacLab
Unified framework for robot learning built on NVIDIA Isaac Sim
theap06/verifiers
Our library for RL environments + evals
theap06/homework_spring2026
theap06/FastVideo
A unified inference and post-training framework for accelerated video generation.
theap06/mlrbench
[NeurIPS 2025 D&B Track] MLR-Bench: Evaluating AI Agents on Open-Ended Machine Learning Research
theap06/SkyRL
SkyRL: A Modular Full-stack RL Library for LLMs
theap06/harness-localized-rl
Harness-localized credit assignment and certificate-guided RL for AppWorld agents
theap06/LMCache
LMCache: Supercharge Your LLM with the Fastest KV Cache Layer
theap06/harbor-datasets
theap06/terminal-bench-3
🚧 Accepting Task Submissions 🚧
theap06/gepa
Optimize prompts, code, and more with AI-powered Reflective Text Evolution
theap06/autoresearch-ratchet-redteam
theap06/harbor
Harbor is a framework for running agent evaluations and creating and using RL environments.
theap06/llm-compressor
Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM
theap06/marin
Open-source framework for the research and development of foundation models.
theap06/gcm
GPU Cluster Monitoring (GCM): Large-Scale AI Research Cluster Monitoring
theap06/pytorch
Tensors and Dynamic neural networks in Python with strong GPU acceleration
theap06/llm-trace
theap06/trackle
theap06/auto_attn
theap06/rl
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.