ByteDance released the Cola2 language model
The story so far
ByteDance has released the continuous latent language model Cola2, which combines dynamic compression VAE, block causal diffusion Transformer (DiT), and cross-segment causal token decoder. It dynamically groups text into semantic segments, uses a compact continuous latent representation for each segment, generates latent blocks, and implements them as discrete tokens. Training code is not included in this release.
AI-written from the reports · Updated 12 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- ByteDance Seed · Hugging Face new modelsByteDance-Seed/Cola2
Cola2 is a continuous latent language model that combines dynamic compression VAE, block causal diffusion Transformer (DiT), and cross-segment causal token decoder, dynamically grouping text into semantic segments, representing each segment with a compact continuous latent, generating latent blocks and implementing them as discrete tokens.