Ollaya package for convaiinnovations/laya

README.md

@@ -0,0 +1,51 @@
+---
+license: apache-2.0
+base_model:
+- convaiinnovations/laya
+library_name: onnx
+tags:
+- ollaya
+- onnx
+- decision-model
+- system-one
+pipeline_tag: text-classification
+---
+
+# laya for Ollaya
+
+[Ollaya](https://github.com/ollaya-dev/ollaya) package of **[convaiinnovations/laya](https://huggingface.co/convaiinnovations/laya)** by Convai Innovations.
+Ollaya runs open decision models locally, the way Ollama runs LLMs: typed questions in,
+calibrated answers out, behind a TypeSafe-compatible API.
+
+```sh
+ollaya run laya
+```
+
+## What is in this repository
+
+This repository holds only the files Ollaya derives, with no weights. Each graph is an ONNX export of the
+original model whose weights **reference the author's own `model.safetensors` by byte offset**,
+so `ollaya pull` downloads the weights from the upstream repository, unmodified and pinned to a
+commit, and verifies their sha256.
+
+| Tag | Upstream | Files |
+|---|---|---|
+| `laya:en` | [convaiinnovations/laya@aa8c91c](https://huggingface.co/convaiinnovations/laya/tree/aa8c91ca088ec597df95a0d1c76b3063cb2ae5e8) | `en/model-fp32.onnx`, `en/model-fp16.onnx`, `en/decision.json`, `en/calibration.json` |
+| `laya:multilingual` | [convaiinnovations/laya@aa8c91c](https://huggingface.co/convaiinnovations/laya/tree/aa8c91ca088ec597df95a0d1c76b3063cb2ae5e8) | `multilingual/model-fp32.onnx`, `multilingual/model-fp16.onnx`, `multilingual/decision.json`, `multilingual/calibration.json` |
+| `laya:typed-decisions` | [convaiinnovations/laya@aa8c91c](https://huggingface.co/convaiinnovations/laya/tree/aa8c91ca088ec597df95a0d1c76b3063cb2ae5e8) | `typed-decisions/model-fp32.onnx`, `typed-decisions/model-fp16.onnx`, `typed-decisions/decision.json`, `typed-decisions/calibration.json` |
+
+Each tag has an fp32 graph (CPU) and an fp16 graph (GPU). Each tag also has `decision.json` (sequence layout, special tokens) and
+`calibration.json` (temperatures).
+
+## Parity
+
+The exports are checked against the PyTorch reference on 2,383 questions per checkpoint.
+The checks use typed-decisions plus multilingual and edge cases:
+
+- **fp32:** the same decision on 100% of questions, with probabilities within 1.1e-4.
+- **fp16:** the same decision on 99.1–99.6% of questions. Nearly all of the differences are
+  near-ties between the top two options.
+
+## License
+
+Same as the upstream model (Apache-2.0). Ollaya itself is Apache-2.0.

en/calibration.json

@@ -0,0 +1,15 @@
+{
+  "temperature": [
+    1.6369030475616455,
+    1.2514300346374512,
+    1.983399510383606
+  ],
+  "temperature_by_options": {
+    "choice:3-5": 1.7601518630981445,
+    "choice:6-10": 1.0000158548355103,
+    "score:3-5": 1.2514300346374512,
+    "noul:2": 1.983399510383606,
+    "choice:11+": 0.10058280825614929,
+    "choice:2": 1.9063563346862793
+  }
+}

en/decision.json

@@ -0,0 +1,28 @@
+{
+  "engine": "onnx",
+  "family": "laya",
+  "layout": "laya-markers-v1",
+  "encoder": "answerdotai/ModernBERT-large",
+  "max_len": 512,
+  "head_max_len": 192,
+  "special_tokens": {
+    "cls": 50281,
+    "sep": 50282,
+    "mask": 50284,
+    "pad": 50283,
+    "mask_text": "[MASK]"
+  },
+  "inputs": [
+    "input_ids",
+    "attention_mask",
+    "marker_pos",
+    "marker_mask",
+    "qtype"
+  ],
+  "outputs": [
+    "logits",
+    "act_logits"
+  ],
+  "min_markers": 2,
+  "opset": 20
+}

en/model-fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32b6a61761499a64d8f20406ed1baa5fe6f0e860a65717c03f804b15bc958886
+size 3708204

en/model-fp32.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c54880e1ab520b5f99ef8417ed2a1f7ef19414e90e87681d1497b39147f8f97a
+size 3722668

multilingual/calibration.json

@@ -0,0 +1,8 @@
+{
+  "temperature": [
+    1.0,
+    1.0,
+    1.0
+  ],
+  "temperature_by_options": {}
+}

multilingual/decision.json

@@ -0,0 +1,28 @@
+{
+  "engine": "onnx",
+  "family": "laya",
+  "layout": "laya-markers-v1",
+  "encoder": "jhu-clsp/mmBERT-base",
+  "max_len": 1024,
+  "head_max_len": 256,
+  "special_tokens": {
+    "cls": 2,
+    "sep": 1,
+    "mask": 4,
+    "pad": 0,
+    "mask_text": "<mask>"
+  },
+  "inputs": [
+    "input_ids",
+    "attention_mask",
+    "marker_pos",
+    "marker_mask",
+    "qtype"
+  ],
+  "outputs": [
+    "logits",
+    "act_logits"
+  ],
+  "min_markers": 2,
+  "opset": 20
+}

multilingual/model-fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:236fbdec33bd2aa44ee4dae43a83173ef3351c4eed6afd073df387c9486ec7d6
+size 2969921

multilingual/model-fp32.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4437087513397063f5d53a904a2404c4bda3cbf4ab30b3c4813678ea9f49e8e
+size 2981576

typed-decisions/calibration.json

@@ -0,0 +1,15 @@
+{
+  "temperature": [
+    1.0148024559020996,
+    1.0374259948730469,
+    1.0575125217437744
+  ],
+  "temperature_by_options": {
+    "choice:3-5": 1.7601518630981445,
+    "choice:6-10": 1.0000158548355103,
+    "score:3-5": 1.2514300346374512,
+    "noul:2": 1.983399510383606,
+    "choice:11+": 0.10058280825614929,
+    "choice:2": 1.9063563346862793
+  }
+}

typed-decisions/decision.json

@@ -0,0 +1,28 @@
+{
+  "engine": "onnx",
+  "family": "laya",
+  "layout": "laya-markers-v1",
+  "encoder": "answerdotai/ModernBERT-large",
+  "max_len": 1024,
+  "head_max_len": 256,
+  "special_tokens": {
+    "cls": 50281,
+    "sep": 50282,
+    "mask": 50284,
+    "pad": 50283,
+    "mask_text": "[MASK]"
+  },
+  "inputs": [
+    "input_ids",
+    "attention_mask",
+    "marker_pos",
+    "marker_mask",
+    "qtype"
+  ],
+  "outputs": [
+    "logits",
+    "act_logits"
+  ],
+  "min_markers": 2,
+  "opset": 20
+}

typed-decisions/model-fp16.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4e9008a387a95b3c164f8ca3bc3c96bbdd2074f1a91b2d51c952a3b6b5d38ab
+size 3600993

typed-decisions/model-fp32.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d648ef513e8d9f1d66a98ccb1d0a71b6902cb776d5332e90b56cea10c1b6e1e
+size 3722669