SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI developer platforms

3,662 results · Page 43 of 102

stable-worldmodel

AI developer platforms

stable-worldmodel is a platform for reproducible world model research and evaluation, developed by galilai-group. It provides a unified interface covering the three stages of world model research: data collection, training, and evaluation based on model predictive control (MPC), and supports a large number of standardized environments. The platform includes reference implementations of common baseline models and planning solvers, allowing research code to focus on the models and objectives themselves. On the data side, the default dataset format is based on the Lance stack (lancedb, pylance, pyarrow), provided separately as an optional dependency; it also supports LeRobot datasets (requires Python 3.12+, installed via pip install 'stable-worldmodel[lerobot]'). Datasets and checkpoints are stored by default in $STABLEWM_HOME (default ~/.stable_worldmodel/) and can be customized. Environments come from the DeepMind Control Suite, Gymnasium classic control, OGBench, Craftax, the Arcade Learning Environment (100+ Atari games), and classic world model benchmarks (Two-Room, PushT). Most environments include independently controllable visual and physical variation factors, making it easy to evaluate zero-shot generalization. After installation, the swm command-line tool is provided for inspecting/converting datasets, environments, and checkpoints. The library is under active development, and the API may change between minor versions.

Read more

SteptronOss

AI developer platforms

SteptronOss is a lightweight, AI-native large language model training framework open-sourced by stepfun-ai, aimed at rapid iteration, reproducible experiments, and modular configuration, covering SFT, RLVR, and evaluation workflows. It can run with only PyTorch as a dependency, while also supporting operator-level replacement for acceleration. Features include configuration-driven experiments with dynamic validation (cfshow, sanity_check), multi-task orchestration with flexible launch tools, and an extensible data/optimizer/model stack. It is suitable for researchers and engineers conducting large model training, fine-tuning, and reinforcement learning experiments.

Read more

storm

AI developer platforms

STORM is an LLM-based knowledge organization system developed by the Stanford OVAL team that can conduct web research on a topic and generate long-form reports with citations. The system consists of two stages, "Research and Outline" and "Writing," and enhances the depth and breadth of questions through multi-perspective questioning and simulated dialogue; its collaborative version, Co-STORM, introduces a human-machine collaborative discussion protocol, multiple LLM agents (experts, moderators), and a dynamic mind map. Users can install it via pip install knowledge-storm, and it supports various language models and embedding models from litellm, as well as multiple retrieval sources such as You.com, Bing, Serper, Brave, DuckDuckGo, and Tavily, and also supports VectorRM based on self-built documents. It is suitable for knowledge exploration and material organization in the early stages of writing, but the official documentation states that its output cannot yet be used directly as a publishable article.

Read more