Compact QEV research direction

The size requirement

The released QEV checkpoint retains a 2.213B-parameter inference backbone. Its 7.992M stored adaptation/readout parameters are not its full model size. The next architecture should target 300-450M total inference parameters, including vision, text, fusion and decision readouts. This is a proposed engineering budget, not an achieved model.

The LAYA model card reports 421M parameters for its ModernBERT-large English model and 322M for its multilingual model. Both are text models. Their size follows from their chosen encoder architecture. LAYA defines RLCD as Reinforcement Learning for Calibrated Decisions, optimizing probability reports with proper scoring rules. That training objective does not remove layers or parameters.

Architecture proposal

Use a compact pretrained text encoder and a compact vision encoder, connected through one trainable fusion and decision network. Candidate descriptions remain part of the request. The head scores candidate-conditioned evidence rather than a fixed label list. The model supports text-only requests and image-plus-text requests in the same network.

Allocate the parameter budget across both encoders and fusion before selecting weights. Avoid retaining a 1.88B language decoder simply because its output generation is unused. A caption-then-text pipeline would add an error-prone intermediary representation and does not meet the goal of direct integrated visual decisions.

Training sequence

Establish supervised text, image and image-plus-policy performance in the compact architecture using licensed data and human or deterministic ground truth.
Distill useful probability distributions and representations from larger teachers, retaining ground-truth supervision. QEV's documented visual and uncertainty failures must not become unquestioned teacher targets.
Compare RLCD-style probability training against cross-entropy, Brier and ordinal proper-scoring objectives under matched data and compute budgets. RLCD is a proposed experiment, not a claimed feature of the current released checkpoint.
Calibrate and select abstention on a separate split. Quantize only after a reliable smaller architecture is established.

Quantization reduces bytes per parameter, not the number of parameters. Structured pruning can reduce dimensions, but requires recovery training and accuracy checks. Distillation trains a genuinely smaller student network; its quality is not guaranteed.

Release gates

Count all parameters, model files and peak serving memory; report merged and unmerged sizes separately where adapters exist.
Preserve request-defined candidates and choice/score/noul behavior.
Use disjoint source, image, group and task-template splits, with untouched final tests.
Report accuracy, distribution error, NLL, ECE, ordinal MAE and selective risk/coverage.
Measure text and image latency on the same RTX 4060 Ti workload and include cold start and concurrency separately. Retain the earlier 300 ms target as a measurement goal.
Test candidate permutations, unseen policies and visual failures rather than relying on narrow training-family improvements.

Implementation and training of this compact architecture are separate from the QEV 0.1.1 naming and Python-package release. No compact weights have been produced yet.