awesome-repositories.com
Blog
MCP
awesome-repositories.com

Descoperă cele mai bune repository-uri open source cu căutare AI.

ExploreazăCăutări recomandateAlternative open-sourceSoftware self-hostedBlogHartă site
ProiectDespreCum realizăm clasamentulPresăServer MCP
LegalConfidențialitateTermeni
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
·
openai avatar

openai/Video-Pre-Training

0
View on GitHub↗
1,712 stele·171 fork-uri·Python·MIT·13 vizualizări

Video Pre Training

Video-Pre-Training este un framework de machine learning conceput pentru antrenarea agenților autonomi în vederea efectuării sarcinilor complexe prin observarea și imitarea comportamentului uman din înregistrări video. Oferă un set de instrumente cuprinzător pentru cercetarea în domeniul învățării prin imitație și învățării prin consolidare (reinforcement learning), permițând dezvoltarea agenților care pot replica acțiunile umane în medii digitale simulate.

Framework-ul se distinge prin capacitatea sa de a procesa seturi de date video la scară largă, neetichetate, pentru a bootstrap-a capabilitățile agenților. Utilizează modelarea dinamicii inverse pentru a deduce input-urile de control din tranzițiile cadrelor, permițând sistemului să mapeze observațiile vizuale la secvențe de acțiune specifice chiar și atunci când etichetele explicite de acțiune lipsesc. Acest proces este susținut de modelarea secvențelor temporale și reprezentarea trăsăturilor latente, care captează relațiile cauzale și dependențele pe termen lung necesare pentru executarea sarcinilor în mai mulți pași.

Dincolo de învățarea inițială prin imitație, platforma include pipeline-uri pentru clonarea comportamentală și optimizarea prin reinforcement learning. Aceste instrumente permit cercetătorilor să rafineze modelele pre-antrenate prin interacțiune iterativă cu mediul și feedback bazat pe recompense. Framework-ul oferă, de asemenea, utilitare pentru capturarea demonstrațiilor umane și standardizarea segmentelor video brute și a log-urilor de acțiune în seturi de date structurate, potrivite pentru ingestia în modele.

Features

  • Behavioral Cloning Toolkits - A collection of tools for mapping visual input from video demonstrations to specific control commands for agent policy development.
  • Imitation and Reinforcement Learning Toolkits - Provides a comprehensive toolkit for training autonomous agents through both imitation learning from video and reinforcement learning optimization.
  • Imitation Learning Pipelines - Trains neural network policies by mapping raw visual observations to discrete action sequences derived from human gameplay demonstrations.
  • Inverse Dynamics Models - Predicts missing action labels from unlabeled video data by training a secondary model to infer control inputs from frame transitions.
  • Reinforcement Learning Research Frameworks - Offers a research-oriented framework for refining pre-trained agent models using iterative environment interaction and reward-based feedback.
  • Imitation Learning Trainings - Establishes foundational agent policies by observing human video demonstrations using behavioral cloning techniques.
  • Video-Based Behavior Replicators - Enables agents to replicate human actions by identifying and mimicking specific behaviors from video input.
  • Agent Demonstration Recorders - Records screen observations, input actions, and environment state snapshots for creating structured training datasets.
  • Autonomous Agent Simulations - Builds intelligent systems that learn to navigate and complete objectives in simulated environments by watching unlabeled online video demonstrations.
  • Action Recognition Systems - Analyzes video frames to infer and replicate human player inputs within simulated digital environments.
  • Imitation Learning Bootstrapping - Utilizes massive datasets of human video recordings to bootstrap agent capabilities before fine-tuning with reinforcement learning.
  • Video Action Inference - Analyzes video recordings to map visual input frames to corresponding sequences of control commands for agent execution.
  • Observation-to-Action Mappings - Translates continuous visual input into structured command outputs by aligning observed human movements with the agent control interface.
  • Reinforcement Learning Optimizers - Refines pre-trained models through iterative environment interaction and reward-based feedback for accelerated goal achievement.
  • Video-Based Task Learning - Allows agents to acquire complex skills by observing and imitating human actions from large unlabeled video datasets.
  • Temporal Sequence Processors - Processes consecutive video frames to capture the causal relationships and long-term dependencies required for complex task execution.
  • Model Fine-Tuning - Adapts pre-trained agent models for specific tasks using custom video demonstrations and action logs.
  • Input-to-Action Mappings - Maps visual input from video recordings to specific control commands to enable agents to replicate human actions in digital environments.

Istoric stele

Graficul istoricului de stele pentru openai/video-pre-trainingGraficul istoricului de stele pentru openai/video-pre-training

Căutare AI

Explorează mai multe repository-uri excelente

Descrie ce ai nevoie în limbaj simplu — AI-ul sortează mii de proiecte open source selectate în funcție de relevanță.

Start searching with AI

Colecții curatoriate care includ Video Pre Training

Colecții selectate manual în care apare Video Pre Training.
  • o bibliotecă pentru construirea modelelor de deep learning

Alternative open-source pentru Video Pre Training

Proiecte open-source similare, clasificate după numărul de funcționalități comune cu Video Pre Training.
  • markfzp/mobile-alohaAvatar MarkFzp

    MarkFzp/mobile-aloha

    4,440Vezi pe GitHub↗

    Mobile Alohas is a bimanual mobile manipulation framework designed to learn whole-body robot control. It provides an integrated imitation learning pipeline that manages the process of collecting human demonstration data and training behavior models to automate complex physical tasks. The system features a robotic teleoperation interface that maps human movements to a mobile robot with dual arms. It includes a whole-body motion dataset tool used for recording, visualizing, and replaying joint and sensor data from manipulation sessions. The framework covers several capability areas, including

    Jupyter Notebookimitation-learningrobotics
    Vezi pe GitHub↗4,440
  • real-stanford/diffusion_policyAvatar real-stanford

    real-stanford/diffusion_policy

    4,299Vezi pe GitHub↗

    Diffusion Policy is a robot learning framework that uses diffusion models to map visual observations to precise action trajectories. It functions as an imitation learning toolkit and visuomotor policy learner, providing a system to train neural networks that replicate human behavior by generating robotic movements based on image and sensor data. The framework employs a conditional denoising process to sample sequences of robotic movements, allowing it to handle multimodal action distributions where multiple valid trajectories may exist for a single state. It utilizes score-based action modeli

    Pythonrobotics
    Vezi pe GitHub↗4,299
  • facebookresearch/reagentAvatar facebookresearch

    facebookresearch/ReAgent

    3,703Vezi pe GitHub↗

    ReAgent is a reinforcement learning platform designed for training, deploying, and evaluating reinforcement learning models and contextual bandit systems for large-scale decision making. It provides a comprehensive suite of tools that spans the entire workflow from initial feasibility analysis to production serving. The system includes a deep reinforcement learning training framework for distributed off-policy algorithms and a specialized model serving layer for high-volume production inference. It distinguishes itself with a counterfactual policy evaluator for estimating performance using hi

    Python
    Vezi pe GitHub↗3,703
  • deepmind/labAvatar deepmind

    deepmind/lab

    7,365Vezi pe GitHub↗

    Lab is a customizable 3D platform and research testbed designed for training and testing autonomous agents using reinforcement learning. It serves as a spatial AI training simulator where agents can be evaluated through navigation and puzzle-solving tasks. The environment allows for the definition of complex layouts and task behaviors through external scripting, enabling the generation of specific challenges for AI research. It supports both automated training via standard API bindings and manual agent control to validate simulation dynamics. The system utilizes a grid-based spatial represen

    C
    Vezi pe GitHub↗7,365
Vezi toate cele 30 alternative pentru Video Pre Training→

Întrebări frecvente

Ce face openai/video-pre-training?

Video-Pre-Training este un framework de machine learning conceput pentru antrenarea agenților autonomi în vederea efectuării sarcinilor complexe prin observarea și imitarea comportamentului uman din înregistrări video. Oferă un set de instrumente cuprinzător pentru cercetarea în domeniul învățării prin imitație și învățării prin consolidare (reinforcement learning), permițând dezvoltarea agenților care pot replica acțiunile umane în medii digitale simulate.

Care sunt principalele funcționalități ale openai/video-pre-training?

Principalele funcționalități ale openai/video-pre-training sunt: Behavioral Cloning Toolkits, Imitation and Reinforcement Learning Toolkits, Imitation Learning Pipelines, Inverse Dynamics Models, Reinforcement Learning Research Frameworks, Imitation Learning Trainings, Video-Based Behavior Replicators, Agent Demonstration Recorders.

Care sunt câteva alternative open-source pentru openai/video-pre-training?

Alternativele open-source pentru openai/video-pre-training includ: markfzp/mobile-aloha — Mobile Alohas is a bimanual mobile manipulation framework designed to learn whole-body robot control. It provides an… real-stanford/diffusion_policy — Diffusion Policy is a robot learning framework that uses diffusion models to map visual observations to precise action… facebookresearch/reagent — ReAgent is a reinforcement learning platform designed for training, deploying, and evaluating reinforcement learning… deepmind/lab — Lab is a customizable 3D platform and research testbed designed for training and testing autonomous agents using… open-gigaai/giga-brain-0 — giga-brain-0 is a robot action model framework designed to train and deploy neural networks that map multi-modal… kenshohara/3d-resnets-pytorch — This project is a PyTorch implementation of 3D residual networks designed for video action recognition. It provides a…