Using INT2 for Coarse Prediction Within a Stabilized FP4/FP8 Substrate (Native Hopper Implementation)

I recently published an architectural research note on using INT2 not as

a replacement precision, but as a predictive head nested inside an

FP8/packed-FP4 precision hierarchy.

## Hypothesis

A stabilized FP4 representation can provide a bounded target surface on

which an INT2 predictor learns a coarse estimate, leaving the full model

body to compute only the residual that the predictor cannot reliably

supply.

The native C++/CUDA implementation is currently training on Hopper.

Because Hopper does not natively execute the complete intended INT2/FP4

compute hierarchy, the low-bit representations are software-managed and

expanded into supported compute formats. The current work therefore

focuses on representational learnability and the control model:

1. Preserving mutable delayed-scaling state across activation-checkpoint

recomputation.

2. Enforcing per-layer ownership of packed-FP4 state histories.

3. Training an INT2 coarse predictor inside that stabilized envelope.

4. Measuring future delivered coverage rather than trusting an

in-window confidence value.

5. Using a governor to grant limited and reversible authority only after

external calibration.

6. Measuring actual residual-work reduction before making a throughput

claim.

The broader idea is to treat extreme low precision as a division of

labor—coarse prediction versus residual correction—rather than as

uniform compression of the full operation.

The article includes the low-bit state invariants, the Lion learning-rate

correction for the predictive head, the nested INT2 → FP4 → FP8 model,

and explicit falsification gates.

**Full research note:**

[From FP4 Stability to INT2 Prediction]( From FP4 Stability to INT2 Prediction )

I am particularly interested in how others are handling temporal

quantizer-state mutation inside activation-checkpoint recomputation in

native CUDA training systems.

Could you explain the purpose of the posts?

You found something out and make blog entries and give the same information (= cross post) or a summary in the forum?

You have additional questions? Request feedback?

It would be good to format the forum entries correctly instead of just pasting markdown format as text.

(In my opinion a hand-written summary would also help to understand, what you or the AI did. Otherwise readers just post into their AI.)

If the purpose is to just get more clicks or links to the (AI generated) blog posts, the posts would have to be moderated.