The story so far
Weights, every dated article, and repo commits on one rail.
13 repo commits since the weights landed
Teal is the weights (2026-09-18), grey ticks are repo commits.
laya-typed-decisions
convaiinnovations Apache 2.0 out 37 days ago
421M parameters, text in, labels out.
Verified 2026-10-25 Rank #0 in Decision models · score 0 method v1 Sources 3 Downloads 30d 3.8k · #558 of 619 tracked
Changed this week: 2 new quantized builds on the Hub (16 in all).2026-10-24
Do people use it?3.8kdownloads, 30 days#558 of 619 we track · below the median+2 likes, downloads +36% this week166 likesHugging Face2026-10-24
Where it's used
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
No API provider serves it, so its downloads are people running it themselves: 16 quantized builds and 7 Spaces.
- As a service0no API provider listed
- On own hardware16quantized builds
- Built on1313 fine-tunes · 0 adapters · 0 merges
- In apps7Hugging Face Spaces
Top Spaces running it: hugging-apps/laya-hebrew ♥ 1
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 0.3 GB at INT4, 0.5 GB at FP8, 1.0 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
29 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
29 derivatives on the Hub, most of them quantized
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
Benchmarks
15 rows
Benchmarks 15 rows
measured third-party leaderboard rowauthor-reported number from the model card or eval files (15)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported Brier | 0.06 | — | model author | 2026-10-07 |
| author-reported ECE | 0.21 | — | model author | 2026-10-07 |
| author-reported ECE | 0.19 | noul | model author | 2026-10-07 |
| author-reported ECE | 0.26 | choice | model author | 2026-10-07 |
| author-reported ECE | 0.2 | score | model author | 2026-10-07 |
| author-reported accuracy | 0.77 | — | model author | 2026-10-07 |
| author-reported accuracy | 0.8 | invoice processing | model author | 2026-10-07 |
| author-reported accuracy | 0.77 | security incidents | model author | 2026-10-07 |
| author-reported accuracy | 0.76 | customer service | model author | 2026-10-07 |
| author-reported accuracy | 0.73 | agent-trace observability | model author | 2026-10-07 |
| author-reported accuracy | 0.86 | noul | model author | 2026-10-07 |
| author-reported accuracy | 0.73 | choice | model author | 2026-10-07 |
| author-reported accuracy | 0.72 | score | model author | 2026-10-07 |
| author-reported score MAE | 0.24 | — | model author | 2026-10-07 |
| author-reported soft acc | 0.47 | — | model author | 2026-10-07 |
Use-case evidence
2 use cases
What it is good at 2 use cases
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Facts
17 fields
Listed in Decision models
Facts
- laya-typed-decisions has 421M parameters. (safetensors metadata, observed 2026-10-24)
- License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed 2026-10-24)
- Created 2026-09-18, last modified 2026-10-03. (HF API, observed 2026-10-24)
| Pipeline tag | text-classification |
|---|---|
| Library | transformers |
| Architectures | LayaTypedDecisions |
| Model type | laya |
| Params by dtype | F16: 421,293,830 |
| Size band | S |
| MoE | no |
| Languages | en |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: not in registry · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: not in registry · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Hardware fit
3 estimates
Hardware fit ESTIMATE
Weights-only memory for laya-typed-decisions: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 1.0 GB · Edge (<4 GB) |
|---|---|
| FP8 (1 byte/param) | 0.5 GB · Edge (<4 GB) |
| INT4 (0.5 bytes/param) | 0.3 GB · Edge (<4 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
Decision facts
catalog
Decision-model catalog curated
| How it was made | fine-tune of Laya on LocalLLaMA/typed-decisions train split (teacher-labelled); ~4-5 h on Kaggle free 2xT4 |
|---|---|
| Lineage bucket | F - fine-tune of an open model (public/synthetic data) |
| Lineage evidence | model card |
| Base model | Laya (ModernBERT-large) |
| Hosted price (input) | self-hosted/M |
| Source | https://huggingface.co/convaiinnovations/laya-typed-decisions |
All variants
29 variants
Lineage
quantized: 16finetune: 13
Top variants (29 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| mys/laya-typed-decisions-GGUF | quantized | GGUF | — | — | apache-2.0 | mys | 6 | 3.5k |
| fr0stbit3/laya-typed-decisions-gguf | quantized | gguf | — | — | apache-2.0 | fr0stbit3 | 0 | 917 |
| onnx-community/laya-typed-decisions-ONNX | quantized | — | — | — | apache-2.0 | onnx-community | 0 | 317 |
| aac6fef/laya-typed-decisions-mlx | finetune | mlx | 421M | S | apache-2.0 | aac6fef | 6 | 0 |
| ti3x-m/laya-typed-decisions-onnx | quantized | — | — | — | apache-2.0 | ti3x-m | 1 | 237 |
| Suarify/laya-typed-decisions-onnx | quantized | — | — | — | apache-2.0 | Suarify | 0 | 195 |
| aac6fef/laya-typed-decisions-coreml | quantized | — | — | — | apache-2.0 | aac6fef | 0 | 162 |
| PythiaFinance/8ball | finetune | — | 421M | S | apache-2.0 | PythiaFinance | 2 | 20 |
| rupeshs/laya-typed-decisions-ov-int8 | quantized | int8 | — | — | apache-2.0 | rupeshs | 2 | 0 |
| NMTZ/laya-typed-decisions-421m-coreml-ane | quantized | — | — | — | apache-2.0 | NMTZ | 1 | 32 |
| ahmedandaloes/laya-typed-decisions-CoreAI | quantized | — | — | — | — | ahmedandaloes | 0 | 79 |
| MissingPackage/jev-d-0.4b-onnx | quantized | — | — | — | apache-2.0 | MissingPackage | 1 | 0 |
| VishalMysore/layaForWebTrained | quantized | — | — | — | apache-2.0 | VishalMysore | 1 | 0 |
| aac6fef/laya-typed-decisions-coreml-adaptive-fp16 | quantized | -fp16 | — | — | apache-2.0 | aac6fef | 0 | 45 |
| daviestechlabs/laya-homelab-router | finetune | — | 421M | S | apache-2.0 | daviestechlabs | 0 | 39 |
| vtava/Laya-1.2-speedup-standalone | finetune | — | 421M | S | — | vtava | 0 | 26 |
| vtava/Laya-1.2-speedup-standalone-fp16 | finetune | -fp16 | 421M | S | — | vtava | 0 | 25 |
| aac6fef/laya-typed-decisions-coreml-adaptive-fp32 | quantized | — | — | — | apache-2.0 | aac6fef | 0 | 22 |
| vtava/Laya-1.2-speedup | finetune | — | — | — | — | vtava | 0 | 19 |
| sumitranjan/Jev-Laya-Guard-Lite | finetune | — | 421M | S | apache-2.0 | sumitranjan | 0 | 6 |
| sivasub987/mandate-1-laya | finetune | — | 421M | S | apache-2.0 | sivasub987 | 0 | 0 |
| yehor-oleksiuk/laya-typed-decisions-onnx | quantized | — | — | — | apache-2.0 | yehor-oleksiuk | 0 | 0 |
| dnagpt/laya-bio-models | finetune | — | — | — | other | dnagpt | 0 | 0 |
| VishalMysore/layaMOE | finetune | — | — | — | apache-2.0 | VishalMysore | 0 | 0 |
| VishalMysore/laya-moe-web | quantized | — | — | — | apache-2.0 | VishalMysore | 0 | 0 |
| jayzou3773/laya-ra2-v33 | finetune | — | 421M | S | apache-2.0 | jayzou3773 | 0 | 0 |
| Dyam/sysone-laya-typed-decisions-onnx | quantized | — | — | — | apache-2.0 | Dyam | 0 | 0 |
| Dipto084/safe_laya | finetune | — | 421M | S | apache-2.0 | Dipto084 | 0 | 0 |
| Aquos06/LayaMacro | finetune | — | — | — | cc-by-nc-sa-4.0 | Aquos06 | 0 | 0 |
Repository changelog
13 commits
Repository changelog 13 commits
| Date | Commit | SHA |
|---|---|---|
| 2026-10-03 | chore: root config.json so the Hub tracks downloads | e929ae5 |
| 2026-09-24 | README: laya 0.3.14 release notes and docs site | 1a793eb |
| 2026-09-24 | README: laya 0.3.13 release notes | 92f2b36 |
| 2026-09-24 | README: laya 0.3.12 release notes | 3279c7d |
| 2026-09-23 | README: laya 0.3.11 release notes | bc76315 |
| 2026-09-23 | Model card: calibration note and faster loading | dd07995 |
| 2026-09-19 | Add the Laya mark to the card; document Router.preload / Router.attach | f9ab0b2 |
| 2026-09-19 | chore: point encoder at the upstream model id instead of a local build path | a398bc6 |
| 2026-09-19 | docs: measured benchmarks, family hub links, Router usage, honest limits | ab82d36 |
| 2026-09-18 | Update README.md | 5ee0035 |
| 2026-09-18 | Laya fine-tuned on typed-decisions: Accuracy 0.766 (beats Jev 0.727) | 843893f |
| 2026-09-18 | Laya fine-tuned on typed-decisions: Accuracy 0.766 (beats Jev 0.727) | a3f934b |
| 2026-09-18 | initial commit | b2dde97 |
Similar models
3 models
Similar models
- laya · 421M same size band (S), same task (text-classification), permissive license
- lavoir · 422M same size band (S), same task (text-classification)
- GLiNER2.5-Decide · 486M same size band (S), same task (text-classification), permissive license
Sources
3 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:43 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:18 UTC · Last verified 2026-10-25
Follow laya-typed-decisions
Watch this model
- /feeds/models/convaiinnovations--laya-typed-decisions.xml(RSS 2.0) · every change to laya-typed-decisions
- /feeds/families/convaiinnovations--laya-typed.xml(RSS 2.0) · every change across the laya-typed family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.