dineshj1

@article{shi2025maestro,
      title={Maestro: Orchestrating Robotics Modules with Vision-Language Models for Zero-Shot Generalist Robots}, 
      author={Junyao Shi and Rujia Yang and Kaitian Chao and Selina Bingqing Wan and Yifei Shao and Jiahui Lei and Jianing Qian and Long Le and Pratik Chaudhari and Kostas Daniilidis and Chuan Wen and Dinesh Jayaraman},
      year={2025},
      abstract={Today's best-explored routes towards generalist robots center on collecting ever larger "observations-in actions-out" robotics datasets to train large end-to-end models, copying a recipe that has worked for vision-language models (VLMs). We pursue a road less traveled: building generalist policies directly around VLMs by augmenting their general capabilities with specific robot capabilities encapsulated in a carefully curated set of perception, planning, and control modules. In Maestro, a VLM coding agent dynamically composes these modules into a programmatic policy for the current task and scenario. Maestro's architectu

dineshj1

# First make sure you pip install google-api-python-client
from apiclient.discovery import build
from httplib2 import Http
from oauth2client import file, client, tools

def store_opts(opt):  # opt is the object returned by argparse
    params = dict(opt.__dict__)
    
   # here I remove parameters I don't want recorded in google sheets
    params.pop('retain_in_test', None)

dineshj1

# Written with help from http://warmspringwinds.github.io/tensorflow/tf-slim
# /2016/12/21/tfrecords-guide/
import os

import ipdb
import numpy as np
import tensorflow as tf
from tqdm import tqdm

tf_dir = './'

dineshj1

# Dinesh Jayaraman

# Based on code by
# Authors: Fabian Pedregosa <fabian.pedregosa@inria.fr>
#          Olivier Grisel <olivier.grisel@ensta.org>
#          Mathieu Blondel <mathieu@mblondel.org>
#          Gael Varoquaux
# License: BSD 3 clause (C) INRIA 2011

print(__doc__)

dineshj1

%demo_textprogressbar
%This a demo for textprogressbar script
textprogressbar('calculating outputs: ');
for i=1:100,
    textprogressbar(i);
    pause(0.1);
end
textprogressbar('done');

dineshj1

import argparse
import time

start_time=time.time();
################## Argument Parsing #####################################
parser=argparse.ArgumentParser();
parser.add_argument('-s','--solver', default='', type=str); # if empty, solver is created, else read
parser.add_argument('-res', '--resume_from', default='', type=str); #if not empty, resumes training from given file
parser.add_argument('-ft', '--finetune_from', default='', type=str);

dineshj1

function submitjobs(nosubmit, numjobs, main)
  try
    xlwrite_path='../data_utils/xlwrite/';
    addpath(xlwrite_path);
    javaaddpath([xlwrite_path 'poi_library/poi-3.8-20120326.jar']);
    javaaddpath([xlwrite_path 'poi_library/poi-ooxml-3.8-20120326.jar']);
    javaaddpath([xlwrite_path 'poi_library/poi-ooxml-schemas-3.8-20120326.jar']);
    javaaddpath([xlwrite_path 'poi_library/xmlbeans-2.3.0.jar']);
    javaaddpath([xlwrite_path 'poi_library/dom4j-1.6.1.jar']);
    javaaddpath([xlwrite_path 'poi_library/stax-api-1.0.1.jar']);

dineshj1

from tsne import bh_sne
import numpy as np
from skimage.transform import resize

from matplotlib import pyplot as plt

def gray_to_color(img):
    if len(img.shape) == 2:
        img = np.dstack((img, img, img))
    return img

dineshj1

  // generate [0..n-1]
  auto seq = [](size_t n) -> std::vector<size_t> {
      std::vector<size_t> v(n);
      for (size_t i=0; i<n; ++i) v[i] = i;
      return v;
  };
  auto index = seq(n);

  // n * n distance matrix
  std::vector<D> dists(n * n);