Issues

• There is a substantial discrepancy between my training outcomes and the official LVE benchmarks you have published

  Doubiiu/CodeTalker#89 · Filter by author zhang666153-wq opened on Aug 7, 2026

• Significant performance gap between re-trained model and official checkpoints on BIWI dataset

  Doubiiu/CodeTalker#88 · Filter by author pepper725 opened on Mar 9, 2026

• stage1 batch_size=1?

  Doubiiu/CodeTalker#87 · Filter by author pepper725 opened on Mar 5, 2026

• Question about LVE definition (paper vs implementation)

  Doubiiu/CodeTalker#86 · Filter by author JHong8 opened on Dec 31, 2025

• Inference custom FLAME model

  Doubiiu/CodeTalker#84 · Filter by author EvdoTheo opened on Apr 1, 2025

• One confusion about FDD calculation

  Doubiiu/CodeTalker#83 · Filter by author SteveSZF opened on Feb 5, 2025

• chinese no good

  Doubiiu/CodeTalker#82 · Filter by author xiaomaofei opened on Nov 29, 2024
  1

• Any progress on making the training batchable?

  Doubiiu/CodeTalker#81 · Filter by author yassineAlouini opened on Nov 23, 2024

• Provide demo output vertices

  Doubiiu/CodeTalker#80 · Filter by author felixtaubner opened on Nov 7, 2024
  1

• training problem about loss

  Doubiiu/CodeTalker#79 · Filter by author renrenzsbbb opened on Sep 24, 2024

• How to plot the analysis graph like the graph in the paper?

  Doubiiu/CodeTalker#78 · Filter by author Tankaishen opened on Sep 19, 2024

• Render Mean and Std.

  Doubiiu/CodeTalker#77 · Filter by author wangryhen opened on Jul 23, 2024

• BIWI Dataset Not Available

  Doubiiu/CodeTalker#75 · Filter by author ManOfStory opened on Jun 19, 2024
  4

• use matlab script to handle the biwi datset

  Doubiiu/CodeTalker#74 · Filter by author KangZW-2001 opened on May 21, 2024

• the input audio file's length limit

  Doubiiu/CodeTalker#73 · Filter by author zk19971101 opened on May 13, 2024
  1

• Is it possible to use facial models trained in Chinese

  Doubiiu/CodeTalker#72 · Filter by author huang-1030 opened on May 9, 2024
  3

• Evaluation Problem: About the Measuring Unit(mm) and testing

  Doubiiu/CodeTalker#70 · Filter by author Tankaishen opened on Apr 1, 2024
  1

• Dont know this technology yet, but Before using it I need to ask a question

  Doubiiu/CodeTalker#69 · Filter by author AIhasArrived opened on Jan 29, 2024
  1

• About the lve.txt and fdd.txt annotation

  Doubiiu/CodeTalker#65 · Filter by author KVBK01 opened on Nov 13, 2023

• The hidden_states during prediction stage is the same when a custom dataset is used

  Doubiiu/CodeTalker#62 · Filter by author Shirley-0708 opened on Oct 25, 2023
  4

• Not converge when larger batch size is used in Stage2

  Doubiiu/CodeTalker#56 · Filter by author NickChang97 opened on Aug 23, 2023
  2

• stage2的训练不收敛

  Doubiiu/CodeTalker#54 · Filter by author drj-123456 opened on Aug 11, 2023
  6

• stage2 training does not converge

  Doubiiu/CodeTalker#53 · Filter by author drj-123456 opened on Aug 10, 2023
  1

• Colab notebook doesn't work

  Doubiiu/CodeTalker#50 · Filter by author epris opened on Aug 1, 2023
  7

• AttributeError: 'tuple' object has no attribute 'transpose'

  Doubiiu/CodeTalker#49 · Filter by author ashutoshbk opened on Aug 1, 2023
  6