thu-ml/RoboticsDiffusionTransformer

Issues

• thanks forHow much graphics memory does fine-tuning approximately require for a graphics card?

  #6 opened 2 months ago by lanlankilkil
  0

• Positional Encoding Questions

  #38 opened a month ago by motor-x
  3

• Wonderful Work!! A few questions regarding modelling and simulation tasks

  #42 opened a month ago by gaodechen
  1

• Fail to read dirty bit

  #29 opened a month ago by wu-yutong-525
  3

• missing 'qpos' in bridgeV2 dataset

  #35 opened a month ago by huangxu1991
  1

• Some question about Finetuning with New Dataset

  #37 opened a month ago by shiqitannn
  2

• Port to LeRobot Dataset v2.0?

  #40 opened a month ago by ivelin
  1

• Have you considered providing a Docker image?

  #34 opened a month ago by soul667
  1

• The role of dirty bit

  #36 opened a month ago by COST-97
  1

• Data cleaning - detecting failed episodes

  #41 opened a month ago by Wonder1905
  1

• May i ask if it deploy on a new robot and fine-tuning it with only 1-5tracks？How is the generalization？

  #33 opened a month ago by lanlankilkil
  1

• Do you have any simulation to recommend?

  #32 opened a month ago by gxy-TJU
  1

• The process of detecting dirty bits in a chunk will not stop.

  #31 opened a month ago by liuzhenO
  1

• QKN and RMSN, which is more effective?

  #39 opened a month ago by jshang-bdai
  1

• t5-xxl is too large; can we replace it with the t5-base?

  #19 opened a month ago by yizt
  3

• Would you like to create a wechat / discord group for RDT1B?

  #2 opened 3 months ago by StarCycle
  5

• Questions about open-source fine-tuning datasets

  #30 opened a month ago by JayceNing
  2

• Pretrain Dataset Question:

  #25 opened 2 months ago by motor-x
  11

• Issue with Mobile Aloha Inference in MuJoCo: Robot Wandering Without Performing Actions

  #24 opened 2 months ago by yongzhengqi
  18

• What are the problems/challenges for this robot manipulation solution: 3D reconstruction then ROS control the arm to take the thing?

  #27 opened a month ago by guotong1988
  3

• I would like to inquire about robot vibrations while watching a video.

  #28 opened a month ago by a510721
  3

• Fine-Tuning in Relative Action Space

  #14 opened 2 months ago by lakomchik
  9

• Confused about normalization

  #23 opened 2 months ago by alik-git
  2

• Mobile ALOHA Dataset Conversion Issues

  #16 opened 2 months ago by WangZhiXiong0
  4

• Applicability of RDT-1B on UR5 Robotic Arm and Fine-Tuning Requirements

  #11 opened 2 months ago by SupritYoung
  1

• Preprocessed dataset

  #17 opened 2 months ago by COST-97
  1

• RuntimeError: mat1 and mat2 must have the same dtype, but got BFloat16 and Float

  #21 opened 2 months ago by liushuailxx
  5

• Some doubts regarding fine-tuning

  #20 opened 2 months ago by Henry-Ellis
  8

• Confusions Regarding Finetuning with New Dataset

  #22 opened 2 months ago by JinmingM
  2

• RuntimeError: mat1 and mat2 must have the same dtype, but got BFloat16 and Float

  #18 opened 2 months ago by liushuailxx
  0

• Fine-tuning the model using data in RLDS format

  #15 opened 2 months ago by prilgrim-whtie
  2

• Dataloder thread issues

  #12 opened 2 months ago by budzianowski
  1

• [Question about Input Configuration in RDT-1b] Why Concatenate Action Noise, Proprio Z, and Frequency C?

  #9 opened 2 months ago by Mark-98
  2

• Nice! How about experimental results in some simulation-based scenes?

  #10 opened 2 months ago by Chocolife-96
  1

• How to run inference of the model with a single image and no proprioception data?

  #5 opened 3 months ago by alik-git
  7

• thanks for your work. How much graphics memory does fine-tuning approximately require for a graphics card?

  #7 opened 2 months ago by lanlankilkil
  2

• Can a single graphics card with 24 GB of video memory load the t5-v1_1-xxl model? Do smaller language models still exist?

  #3 opened 3 months ago by qiynxx
  11

• Fine-tuned RDT checkpoint

  #4 opened 3 months ago by lakomchik
  1

• Nice work! How do you deal with different embodiments when sharing the same action space?

  #1 opened 3 months ago by StarCycle
  1