Open Models Almanac
Methodspost-training

general family · filed under post-training

post-training

Post-training is a second training stage that follows discrete-token pre-training.1 It specializes the model to the target use-case while adding a flow-matching action expert for continuous actions, training jointly with next-token prediction to keep text capabilities.2

Trade-off Jointly training with next-token prediction preserves the model's text prediction capabilities.3

How this definition was written

Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.

  1. After pre-training the model with discrete tokens for 280k gradient steps, we perform a second stage of training that we refer to as post-training — π0.5\pi_{0.5}: a Vision-Language-Action Model with Open-World Generalization
  2. to both specialize the model to our use-case (mobile manipulation in homes), and to add an action expert that can produce continuous action chunks via flow matching — π0.5\pi_{0.5}: a Vision-Language-Action Model with Open-World Generalization
  3. This stage jointly trains with next-token prediction, to preserve text prediction capabilities — π0.5\pi_{0.5}: a Vision-Language-Action Model with Open-World Generalization

Also called Flow-matching action-expert post-training, hybrid training procedure.

sources
3
models
3
labs adopt it
3
strongest
core

Documented in

Allen Institute for AI
Z.ai (Zhipu AI)
Physical Intelligence

Evidence

4 quotes from 3 sources, one card per model, each opening with its strongest.

How sources treat it
used 2core 2

π0.5

Physical Intelligenceadopted1 source

uses a hybrid training procedure, with discrete tokenized training in the pre-training phase, and training with a flow matching action expert only in the post-training phase

usedunclearin π0.5
1 more from this model

training with a flow matching action expert only in the post-training phase

usedunclearin π0.5
↑