jihoonerd / Conditional-Motion-In-Betweening

🕹️ Official Implementation of Conditional Motion In-betweening (CMIB) 🏃
https://jihoonerd.github.io/Conditional-Motion-In-Betweening/
123 stars 9 forks source link

[Major Change] Use BERT-based Transformer Encoder / Transformer Decoder #38

Closed jihoonerd closed 3 years ago

jihoonerd commented 3 years ago

Experiments conducted has shown that finding manifold of motion generation and disentanglement of it is difficult.

BERT suggested not only a transformer based representation learning, but also presented Masked Language Model(MLM) task, which considerably resembles our in-betweening work.

Similar approach is studied at https://arxiv.org/abs/2103.00776.

Any kinds of discriminator (1D Conv, Transformer Decoder, RNN-based Decoder) can be integrated to yield mutual information loss.