laya:typed-decisionstrained on split
Laya · Convai Innovations · 421M · ONNX Runtime · Apache-2.0
- Score
- 40.3
- 95 % interval 37.8–42.6
- Rank
- 21
- of 35 scored models
- Public benchmark
- 59.4
- macro accuracy, %
- typed-decisions
- 76.7
- trained on its train split
By task
Chance-corrected accuracy on each of the 14 tasks: 0 is guessing, 100 is every answer right. * trained on its train split, so not in the score.
Accuracy and calibration
| Benchmark | Accuracy | Choice | Noul | Score | ECE | Brier | Rejected |
|---|---|---|---|---|---|---|---|
| Public benchmark | 59.4(57.9–60.8) | 63.9 | 69.7 | 34.5 | 0.022 | 0.475 | 31 / 3,880 |
| typed-decisions * | 76.7(74.6–78.7) | 73.3 | 85.8 | 72.4 | 0.214 | 0.400 | 0 / 400 requests |
Accuracy in %, with its 95 % bootstrap interval. A rejected request (the model can’t take the question, or the input is too long for it) counts as wrong.
Speed and memory
| Device | p50 | p90 | Load | VRAM | RAM peak | Precision |
|---|---|---|---|---|---|---|
| RTX 5090 CUDA | 20.4 ms | 21.1 ms | 3.86 s | 1.6 GB | 1.3 GB | F16 |
| Threadripper 3960X CPU | 944 ms | 952 ms | 6.84 s | – | 3.1 GB | F32 |
Five-question requests over HTTP, one at a time, each a different short message. VRAM: the rise of the GPU’s used memory while the model is loaded. RAM peak: the runner process’s peak resident memory.
Parity with the authors’ code
Its parity check is in the model’s family documentation in the repository (docs/families/); the compiled parity file does not list it yet.
Reproduce
The tools are in convert/ollaya_convert and need only Python’s standard library. Run them against a local Ollaya 0.13.0 with the same model digest.
ollaya pull laya:typed-decisions
# typed-decisions; export the test split once with
# uv run python -m ollaya_convert.bench_typed export --out typed.jsonl
python -m ollaya_convert.bench_typed run --model laya:typed-decisions \
--url http://127.0.0.1:11435 --data typed.jsonl --out typed/laya-typed-decisions
# Bespoke Labs' public benchmark; build its 13 subsets first
NIMBLE_SRC=nimble python -m ollaya_convert.bench_public run --model laya:typed-decisions \
--url http://127.0.0.1:11435 --data public --out public/laya-typed-decisions
# speed and memory
python -m ollaya_convert.bench_latency --models laya:typed-decisions --memory \
--machine my-machine --device cuda --out latency.json