SuperAPI AI Directory
Partner with us中文AI newsFavorites

AI developer platforms

3,661 results · Page 49 of 102

讯飞星火认知大模型

AI developer platforms

iFlytek Spark Cognitive Large Model is a large model service provided by iFlytek through the iFlytek Open Platform. According to the official page introduction, the platform provides services such as the Spark Large Model, speech recognition, speech synthesis, OCR experience, and AIUI human-computer interaction, can adapt to various business scenarios, and supports both online and private deployment. Users can log in to the platform to claim tens of millions of free Tokens.

Read more

1.5-Pints

AI developer platforms

1.5-Pints is a compact large language model (LLM) project developed by Pints.AI. Its codebase contains the model architecture, training scripts, and utility tools, and also includes the 0.12-Pint model. The project aims to promote model reproduction, experimentation, and further open-source development by publicly releasing the model codebase and architecture. According to the materials, 1.5-Pints is a compact LLM pretrained on high-quality data within 9 days. Users can obtain the code through the GitHub repository and follow the instructions to configure the environment using Ubuntu 22.04 LTS x86-64, a conda environment, the CUDA toolkit, etc., followed by steps such as data preparation, model training, fine-tuning, weight conversion, direct preference optimization (DPO), and evaluation. The project provides a complete training pipeline, including pretraining dataset download, dataset preparation scripts, and training parameter adjustment (such as architecture configuration, micro-batch size, etc.). In addition, the codebase also includes tests, allowing users to run tests after modifications to ensure code stability. The model supports conversion to HuggingFace format (PyTorch and safetensors), and the use of safetensors files is recommended. The project also provides a Discord community link for communication.

Read more

ABot-World

AI developer platforms

ABot-World is a real-time interactive world simulator project for a single desktop GPU, released by the ABot-World Team. Its goal is to achieve infinite, action-conditioned world rollout on a single NVIDIA RTX 5090 desktop graphics card: supporting 720p resolution, 16 FPS, approximately 1.2 seconds of latency, and approximately 19GB of VRAM usage. The project emphasizes action-driven world control, real-time desktop inference, infinite world rollout, and open-ended scene imagination, and adopts a training method called LongForcing to extend new scenes and dynamics during rollout, avoiding scene lock-in and requiring no prompt switching. The repository provides resources such as inference code, a local Gradio demo, an online playground, the causal student model ABot-World-0-5B-LF, a technical report, a bidirectional teacher model, a 500-hour action-annotated video training dataset, and a Docker environment image. The project is licensed under the Apache License 2.0 and is built on top of multiple open-source projects.

Read more

ActWorld

AI developer platforms

ActWorld is a real-time interactive world model that unifies flexible navigation and rich mid-episode object interaction within a single chunked autoregressive framework. Driven by hierarchical action-aware memory and a 100,000-clip interaction-intensive video dataset, it aims to solve the problem that previous navigation-centric world models can only move but cannot truly interact with scene objects. The model alleviates the "action forgetting" pathology through hierarchical action-aware memory, and introduces a persistent memory bank to maintain event updates and object identity markers during long-horizon rollouts. Experiments show that ActWorld simultaneously supports flexible navigation and rich object interaction in a single model, significantly improving interaction fidelity without sacrificing viewpoint control.

Read more

advanced-unity-mcp

AI developer platforms

Advanced Unity MCP is a public repository tool developed by Code Maestro that lets users control the Unity Editor through natural language commands. Users do not need to manually click menus or set up GameObjects; they only need to describe their needs to an AI assistant, and the operations are completed automatically. The tool connects the Unity Editor with AI assistants based on the Model Context Protocol (MCP), and supports the Code Maestro desktop app (direct connection) as well as MCP-capable clients such as GitHub Copilot (VS Code) and Cursor (connected through a relay server). Its main capabilities cover core editor control, asset and scene management, build and platform tools, project configuration, and performance profiling.

Read more