Publish validated native MLX FP16 checkpoint

LICENSE

@@ -0,0 +1,176 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS

NOTICE

@@ -0,0 +1,14 @@
+laya-mlx
+Copyright 2026 laya-mlx contributors
+
+This product includes software derived from Laya:
+https://github.com/NandhaKishorM/laya
+Copyright Convai Innovations and Laya contributors. Licensed under Apache-2.0.
+Upstream source revision: 6a5819129eb220570792e417e49723d697efd76f
+
+The token sequence construction, question rendering, confidence calculation,
+presets, email utilities and language router are adapted from Laya.
+The neural network is reimplemented using Apple's MLX, following Laya's
+DecisionModel and the ModernBERT architecture in Hugging Face Transformers.
+Model weights are downloaded separately from Convai Innovations on Hugging Face;
+they are not included in this repository.

README.md

@@ -0,0 +1,74 @@
+---
+license: apache-2.0
+library_name: mlx
+pipeline_tag: text-classification
+base_model: convaiinnovations/laya
+language:
+- en
+tags:
+- mlx
+- laya
+- modernbert
+- apple-silicon
+- decision-model
+---
+
+# laya-mlx
+
+Native **MLX FP16** conversion of [convaiinnovations/laya](https://huggingface.co/convaiinnovations/laya) for Apple silicon.
+
+This checkpoint uses **ModernBERT-large**, a **512-token total context**, and Laya's decision Transformer, scoring head and action head. It supports `choice`, ordinal `score`, and boolean `noul` questions. All model computation runs in MLX; the runtime does not require PyTorch or Transformers.
+
+## Usage
+
+Install the dedicated runtime on an Apple silicon Mac with macOS 26+ and Python 3.11+:
+
+```bash
+python -m pip install 'git+https://github.com/mizorewww/laya-mlx.git'
+```
+
+```python
+import laya_mlx as laya
+
+agent = laya.load("aac6fef/laya-mlx")
+result = agent.predict(
+    "I was billed twice. Please refund the duplicate today.",
+    {
+        "department": {
+            "type": "choice",
+            "instructions": "Which department should handle this request?",
+            "criteria": ["billing", "technical", "sales"],
+        },
+        "refund": {
+            "type": "noul",
+            "instructions": "Does the customer ask for money back?",
+        },
+    },
+)
+print(result["answers"])
+```
+
+Use `dtype="float32"` for closer agreement with upstream FP32 arithmetic. The source weights themselves are FP16. Question formatting, tokenizer behavior, calibration temperatures and output schema are preserved.
+
+This is a bidirectional decision encoder loaded with `laya_mlx`. The package provides the custom architecture needed to interpret the checkpoint. The repository does not include a generative language model or training implementation.
+
+## Validation
+
+Tested locally on Apple M3 Max, 40-core GPU, 128 GB unified memory, macOS 27.2, Python 3.12.13 and MLX 0.32.2.
+
+- FP16 agrees with upstream PyTorch MPS FP32 on the argmax of **63/63** decision distributions across 16 cases.
+- Maximum calibrated probability difference: **0.0054443**.
+- **100 repeated calls** produced finite, deterministic public outputs; measured MLX active-memory growth after clearing caches was **0 bytes**.
+- Every exported tensor was checked for exact equality with the corresponding source tensor cast to FP16.
+
+The included `validation.json` contains numerical and stability measurements for both FP32 and FP16 arithmetic. [Full performance report and raw timing samples](https://github.com/mizorewww/laya-mlx/blob/main/BENCHMARKS.md) compare MLX with the original runtime on the same machine. These checks establish port fidelity, not that every model answer is correct.
+
+## Provenance and limits
+
+- Source checkpoint: `convaiinnovations/laya` at `c5d78730f3493e4fe16d61507ef4b78eef7318cf`.
+- Upstream code: [NandhaKishorM/laya](https://github.com/NandhaKishorM/laya), commit `6a5819129eb220570792e417e49723d697efd76f`.
+- Conversion changes parameter names for MLX and preserves FP16 weights. It does not retrain or quantize to fewer bits.
+- This is an independent port. Model quality, calibration and language/task limitations remain those of the original checkpoint. Questions and options share the context budget with the input state.
+- The typed-decisions checkpoint is specialized for upstream workflows; the multilingual checkpoint is the intended choice for non-English text.
+
+Apache-2.0. Original Laya models and code are by Convai Innovations and contributors. See `LICENSE`, `NOTICE`, `mlx_config.json` and `manifest.json` for attribution and export details.

encoder/config.json

@@ -0,0 +1,84 @@
+{
+  "architectures": [
+    "ModernBertForMaskedLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 50281,
+  "classifier_activation": "gelu",
+  "classifier_bias": false,
+  "classifier_dropout": 0.0,
+  "classifier_pooling": "mean",
+  "cls_token_id": 50281,
+  "decoder_bias": true,
+  "deterministic_flash_attn": false,
+  "dtype": "float32",
+  "embedding_dropout": 0.0,
+  "eos_token_id": 50282,
+  "global_attn_every_n_layers": 3,
+  "gradient_checkpointing": false,
+  "hidden_activation": "gelu",
+  "hidden_size": 1024,
+  "initializer_cutoff_factor": 2.0,
+  "initializer_range": 0.02,
+  "intermediate_size": 2624,
+  "layer_norm_eps": 1e-05,
+  "layer_types": [
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention",
+    "sliding_attention",
+    "sliding_attention",
+    "full_attention"
+  ],
+  "local_attention": 128,
+  "max_position_embeddings": 8192,
+  "mlp_bias": false,
+  "mlp_dropout": 0.0,
+  "model_type": "modernbert",
+  "norm_bias": false,
+  "norm_eps": 1e-05,
+  "num_attention_heads": 16,
+  "num_hidden_layers": 28,
+  "pad_token_id": 50283,
+  "position_embedding_type": "absolute",
+  "repad_logits_with_grad": false,
+  "rope_parameters": {
+    "full_attention": {
+      "rope_theta": 160000.0,
+      "rope_type": "default"
+    },
+    "sliding_attention": {
+      "rope_theta": 10000.0,
+      "rope_type": "default"
+    }
+  },
+  "sep_token_id": 50282,
+  "sparse_pred_ignore_index": -100,
+  "sparse_prediction": false,
+  "tie_word_embeddings": true,
+  "transformers_version": "5.0.0",
+  "vocab_size": 50368
+}

manifest.json

@@ -0,0 +1,54 @@
+{
+  "repository": "aac6fef/laya-mlx",
+  "source": "convaiinnovations/laya",
+  "source_revision": "c5d78730f3493e4fe16d61507ef4b78eef7318cf",
+  "format": "laya-mlx",
+  "dtype": "float16",
+  "verified_tensors": 206,
+  "files": {
+    "LICENSE": {
+      "bytes": 10173,
+      "sha256": "a6cba85bc92e0cff7a450b1d873c0eaa2e9fc96bf472df0247a26bec77bf3ff9"
+    },
+    "NOTICE": {
+      "bytes": 698,
+      "sha256": "f23ae8ae701f5a43e186182e387aeef00b666daed18a946fc7c9ea827b00a8fa"
+    },
+    "README.md": {
+      "bytes": 3602,
+      "sha256": "68c97e9ef6a45ddeb570f97cecb7f9337280c70d638ddb8701e1fb0f337c8420"
+    },
+    "encoder/config.json": {
+      "bytes": 2083,
+      "sha256": "bf3ab80598fdccf414855a2ce80f22859e4492d06ca8a62ddd1cfb63972f8979"
+    },
+    "mlx_config.json": {
+      "bytes": 290,
+      "sha256": "c022368f128dc9b7f6478a9163c96e0406f1da6a7c6bae660ea8515b016ed7ef"
+    },
+    "model.safetensors": {
+      "bytes": 842609225,
+      "sha256": "b9c07bf14be2fa5c78a9193a3e6d840ac80e89e62fc40f425834c3d8a6eaa3de"
+    },
+    "rl_agent_config.json": {
+      "bytes": 746,
+      "sha256": "d96dc2cb39d6375e030ff48c9957088f3c52668f45c30c56504f4e801ed3ee62"
+    },
+    "tokenizer/tokenizer.json": {
+      "bytes": 3583228,
+      "sha256": "6c8aaa9a542084f2457eab775d4eeb51f92a70c0fd9de28d5edb0ddec3c08d30"
+    },
+    "tokenizer/tokenizer_config.json": {
+      "bytes": 308,
+      "sha256": "50044de60daaa73df97d262e15a40d4faf0160e7d742df64b377877a1320dd12"
+    },
+    "validation.json": {
+      "bytes": 18257,
+      "sha256": "adc9a4baac1430b58cd8fdaf7c7587b380d4a91834ba405bbdeeabd9d35b5771"
+    },
+    ".gitattributes": {
+      "bytes": 1519,
+      "sha256": "11ad7efa24975ee4b0c3c3a38ed18737f0658a5f75a0a96787b576a78a023361"
+    }
+  }
+}

mlx_config.json

@@ -0,0 +1,10 @@
+{
+  "format": "laya-mlx",
+  "format_version": 1,
+  "dtype": "float16",
+  "source": "convaiinnovations/laya",
+  "revision": "c5d78730f3493e4fe16d61507ef4b78eef7318cf",
+  "subfolder": null,
+  "source_revision": "c5d78730f3493e4fe16d61507ef4b78eef7318cf",
+  "repository": "aac6fef/laya-mlx"
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9c07bf14be2fa5c78a9193a3e6d840ac80e89e62fc40f425834c3d8a6eaa3de
+size 842609225

rl_agent_config.json

@@ -0,0 +1,33 @@
+{
+  "encoder": "answerdotai/ModernBERT-large",
+  "head_layers": 2,
+  "max_len": 512,
+  "head_max_len": 192,
+  "max_prefixes": 6,
+  "act_costs": {
+    "escalate": 0.5
+  },
+  "cost_wrong_act": 3.0,
+  "amp_dtype": "bf16",
+  "model_name": "rl-agent",
+  "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
+  },
+  "training": {
+    "updates": 7313,
+    "epochs_completed": 1,
+    "hours": 1.96,
+    "world_size": 1,
+    "fine_tuned_from_checkpoint": true
+  }
+}

tokenizer/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "model_input_names": [
+    "input_ids",
+    "attention_mask"
+  ],
+  "model_max_length": 8192,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "tokenizer_class": "PreTrainedTokenizerFast",
+  "unk_token": "[UNK]"
+}

validation.json

@@ -0,0 +1,462 @@
+{
+  "results": [
+    {
+      "model": "laya",
+      "revision": "c5d78730f3493e4fe16d61507ef4b78eef7318cf",
+      "dtype": "float32",
+      "argmax_agreements": 63,
+      "questions": 63,
+      "probability_max_abs_error": 5.170394474396378e-06,
+      "probability_tolerance": 0.0001,
+      "cases": [
+        {
+          "case": "email",
+          "questions": 3,
+          "sequence_length": 93,
+          "input_sha256": "97703a282ecbb252db9f7e3121ad77e9d1a6671656cf35e62e60f0efeffd032c",
+          "tokens_equal": true,
+          "logits_max_abs_error": 9.387731552124023e-06,
+          "action_logits_max_abs_error": 0.0126953125,
+          "probability_max_abs_error": 1.9275840926757937e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "en",
+          "questions": 3,
+          "sequence_length": 65,
+          "input_sha256": "f89895466f2fafda31ca556a46d3823764e73156772693c88cc148edcc6a39eb",
+          "tokens_equal": true,
+          "logits_max_abs_error": 3.814697265625e-06,
+          "action_logits_max_abs_error": 0.00244140625,
+          "probability_max_abs_error": 5.839881763725074e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "zh",
+          "questions": 3,
+          "sequence_length": 73,
+          "input_sha256": "0872ab22fd69ba60ca320db76dc09d90a9de1f4ebd4aaeafd46169a763f20e3f",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.0251998901367188e-05,
+          "action_logits_max_abs_error": 0.001953125,
+          "probability_max_abs_error": 1.5244581486606457e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "de",
+          "questions": 3,
+          "sequence_length": 76,
+          "input_sha256": "88c970581ad10b6bd1aa82afaaef93513e456e68cf77e7e3cad98919de3ff1a5",
+          "tokens_equal": true,
+          "logits_max_abs_error": 6.67572021484375e-06,
+          "action_logits_max_abs_error": 0.00146484375,
+          "probability_max_abs_error": 2.909347931456807e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "fr",
+          "questions": 3,
+          "sequence_length": 81,
+          "input_sha256": "00d71ea5bfd209fbbeffad3decdfacd9a51a4a97c07b7ab8969af5aa8bd51e43",
+          "tokens_equal": true,
+          "logits_max_abs_error": 6.079673767089844e-06,
+          "action_logits_max_abs_error": 0.014404296875,
+          "probability_max_abs_error": 7.556087676663381e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "es",
+          "questions": 3,
+          "sequence_length": 75,
+          "input_sha256": "3d429a6d7f31bc59256e88c56a1d7a8ae40a05c6277fc6d22341bdc2203d93b6",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.3828277587890625e-05,
+          "action_logits_max_abs_error": 0.00830078125,
+          "probability_max_abs_error": 1.2019799846552104e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "hi",
+          "questions": 3,
+          "sequence_length": 119,
+          "input_sha256": "23f2a78bed3a972ae02a7d3dc26e5b0eec62aee4230ff0c5a1388d7f044dd121",
+          "tokens_equal": true,
+          "logits_max_abs_error": 3.93986701965332e-05,
+          "action_logits_max_abs_error": 0.0205078125,
+          "probability_max_abs_error": 5.170394474396378e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "ja",
+          "questions": 3,
+          "sequence_length": 75,
+          "input_sha256": "e73ae1efad1e50525b04f6a8f8004ccc15478f73ac3618fbe752e25c1c6c150b",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.621246337890625e-05,
+          "action_logits_max_abs_error": 0.0029296875,
+          "probability_max_abs_error": 1.8448868603915258e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "ru",
+          "questions": 3,
+          "sequence_length": 85,
+          "input_sha256": "bac6676f433fdbb4aa2b71dda957299b7e5a9efa2c2b526f000f51d8d4be3a08",
+          "tokens_equal": true,
+          "logits_max_abs_error": 2.2411346435546875e-05,
+          "action_logits_max_abs_error": 0.00341796875,
+          "probability_max_abs_error": 1.3785159388851609e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "empty_state",
+          "questions": 3,
+          "sequence_length": 46,
+          "input_sha256": "85f8a747931b65f225679d74c32afc65b23c679549560503921a969453d2c4e1",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.1205673217773438e-05,
+          "action_logits_max_abs_error": 0.0009765625,
+          "probability_max_abs_error": 8.773777885884293e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "long",
+          "questions": 3,
+          "sequence_length": 512,
+          "input_sha256": "029a742cb223bbb34add55cb46a0d266345a2e448f314ced18033f3710a9c8c7",
+          "tokens_equal": true,
+          "logits_max_abs_error": 6.556510925292969e-06,
+          "action_logits_max_abs_error": 0.00048828125,
+          "probability_max_abs_error": 1.7214299026235835e-06,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "conversation",
+          "questions": 3,
+          "sequence_length": 72,
+          "input_sha256": "a429d0393a0bd2af5d960fa45ab286d6741b89969693fa852560a6c87309a4d0",
+          "tokens_equal": true,
+          "logits_max_abs_error": 8.344650268554688e-06,
+          "action_logits_max_abs_error": 0.00244140625,
+          "probability_max_abs_error": 5.85747966019845e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "mask_literals",
+          "questions": 3,
+          "sequence_length": 55,
+          "input_sha256": "f89667d91cd59690e55b05e6af9368ea1e27ef42829525a97d4cd3cace2f1989",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.1444091796875e-05,
+          "action_logits_max_abs_error": 0.0009765625,
+          "probability_max_abs_error": 9.02355455134618e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "many_questions",
+          "questions": 20,
+          "sequence_length": 93,
+          "input_sha256": "47942741e9ed5fabc4c49464b633b43608c61d04c9bc3f23fea5c0ae4da23ae3",
+          "tokens_equal": true,
+          "logits_max_abs_error": 7.152557373046875e-06,
+          "action_logits_max_abs_error": 0.01220703125,
+          "probability_max_abs_error": 4.862327878663031e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 20,
+          "public_result_equal": true
+        },
+        {
+          "case": "structured",
+          "questions": 3,
+          "sequence_length": 78,
+          "input_sha256": "8ba1aac3447918529ea99ae8c34146ebe16db359e7787412e6dda90c34f9b8b2",
+          "tokens_equal": true,
+          "logits_max_abs_error": 1.1727213859558105e-05,
+          "action_logits_max_abs_error": 0.00146484375,
+          "probability_max_abs_error": 9.181504970756649e-07,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": true
+        },
+        {
+          "case": "twenty_options",
+          "questions": 1,
+          "sequence_length": 137,
+          "input_sha256": "f20fe7490e99ad43d4d1c5a9657f095968b904c9a60e8b3e02fce0544126f838",
+          "tokens_equal": true,
+          "logits_max_abs_error": 2.0503997802734375e-05,
+          "action_logits_max_abs_error": 0.0068359375,
+          "probability_max_abs_error": 6.2494302366671145e-34,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 1,
+          "public_result_equal": true
+        }
+      ],
+      "stability": {
+        "calls": 100,
+        "elapsed_seconds": 3.9830371670013847,
+        "finite": true,
+        "deterministic": true,
+        "active_bytes_before": 1685193880,
+        "active_bytes_after": 1685193880,
+        "active_growth_bytes": 0,
+        "peak_bytes": 3296604064
+      }
+    },
+    {
+      "model": "laya",
+      "revision": "c5d78730f3493e4fe16d61507ef4b78eef7318cf",
+      "dtype": "float16",
+      "argmax_agreements": 63,
+      "questions": 63,
+      "probability_max_abs_error": 0.005444271459395356,
+      "probability_tolerance": 0.02,
+      "cases": [
+        {
+          "case": "email",
+          "questions": 3,
+          "sequence_length": 93,
+          "input_sha256": "97703a282ecbb252db9f7e3121ad77e9d1a6671656cf35e62e60f0efeffd032c",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0036835670471191406,
+          "action_logits_max_abs_error": 1.83251953125,
+          "probability_max_abs_error": 0.0008600417756422507,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "en",
+          "questions": 3,
+          "sequence_length": 65,
+          "input_sha256": "f89895466f2fafda31ca556a46d3823764e73156772693c88cc148edcc6a39eb",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.005011081695556641,
+          "action_logits_max_abs_error": 3.49658203125,
+          "probability_max_abs_error": 0.0003116258116069026,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "zh",
+          "questions": 3,
+          "sequence_length": 73,
+          "input_sha256": "0872ab22fd69ba60ca320db76dc09d90a9de1f4ebd4aaeafd46169a763f20e3f",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.04189777374267578,
+          "action_logits_max_abs_error": 17.02685546875,
+          "probability_max_abs_error": 0.0013265283093753522,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "de",
+          "questions": 3,
+          "sequence_length": 76,
+          "input_sha256": "88c970581ad10b6bd1aa82afaaef93513e456e68cf77e7e3cad98919de3ff1a5",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.005423009395599365,
+          "action_logits_max_abs_error": 0.80078125,
+          "probability_max_abs_error": 0.0002973207056953253,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "fr",
+          "questions": 3,
+          "sequence_length": 81,
+          "input_sha256": "00d71ea5bfd209fbbeffad3decdfacd9a51a4a97c07b7ab8969af5aa8bd51e43",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.004002094268798828,
+          "action_logits_max_abs_error": 5.021240234375,
+          "probability_max_abs_error": 0.0006047629910134744,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "es",
+          "questions": 3,
+          "sequence_length": 75,
+          "input_sha256": "3d429a6d7f31bc59256e88c56a1d7a8ae40a05c6277fc6d22341bdc2203d93b6",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.006280422210693359,
+          "action_logits_max_abs_error": 2.7939453125,
+          "probability_max_abs_error": 0.0015573923004306156,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "hi",
+          "questions": 3,
+          "sequence_length": 119,
+          "input_sha256": "23f2a78bed3a972ae02a7d3dc26e5b0eec62aee4230ff0c5a1388d7f044dd121",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.04454207420349121,
+          "action_logits_max_abs_error": 10.7470703125,
+          "probability_max_abs_error": 0.005444271459395356,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "ja",
+          "questions": 3,
+          "sequence_length": 75,
+          "input_sha256": "e73ae1efad1e50525b04f6a8f8004ccc15478f73ac3618fbe752e25c1c6c150b",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0074710845947265625,
+          "action_logits_max_abs_error": 1.3525390625,
+          "probability_max_abs_error": 0.0009152261989081867,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "ru",
+          "questions": 3,
+          "sequence_length": 85,
+          "input_sha256": "bac6676f433fdbb4aa2b71dda957299b7e5a9efa2c2b526f000f51d8d4be3a08",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.005573272705078125,
+          "action_logits_max_abs_error": 2.29296875,
+          "probability_max_abs_error": 0.000573146121867707,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "empty_state",
+          "questions": 3,
+          "sequence_length": 46,
+          "input_sha256": "85f8a747931b65f225679d74c32afc65b23c679549560503921a969453d2c4e1",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0027773380279541016,
+          "action_logits_max_abs_error": 2.47705078125,
+          "probability_max_abs_error": 0.00039989655194616613,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "long",
+          "questions": 3,
+          "sequence_length": 512,
+          "input_sha256": "029a742cb223bbb34add55cb46a0d266345a2e448f314ced18033f3710a9c8c7",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0040493011474609375,
+          "action_logits_max_abs_error": 2.482421875,
+          "probability_max_abs_error": 0.00038620949119971826,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "conversation",
+          "questions": 3,
+          "sequence_length": 72,
+          "input_sha256": "a429d0393a0bd2af5d960fa45ab286d6741b89969693fa852560a6c87309a4d0",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.004204750061035156,
+          "action_logits_max_abs_error": 2.8369140625,
+          "probability_max_abs_error": 0.001078521392323395,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "mask_literals",
+          "questions": 3,
+          "sequence_length": 55,
+          "input_sha256": "f89667d91cd59690e55b05e6af9368ea1e27ef42829525a97d4cd3cace2f1989",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.020479202270507812,
+          "action_logits_max_abs_error": 2.844482421875,
+          "probability_max_abs_error": 0.0006111239970967852,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "many_questions",
+          "questions": 20,
+          "sequence_length": 93,
+          "input_sha256": "47942741e9ed5fabc4c49464b633b43608c61d04c9bc3f23fea5c0ae4da23ae3",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0018677711486816406,
+          "action_logits_max_abs_error": 5.8330078125,
+          "probability_max_abs_error": 0.00016821344275153027,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 20,
+          "public_result_equal": false
+        },
+        {
+          "case": "structured",
+          "questions": 3,
+          "sequence_length": 78,
+          "input_sha256": "8ba1aac3447918529ea99ae8c34146ebe16db359e7787412e6dda90c34f9b8b2",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.0077322423458099365,
+          "action_logits_max_abs_error": 1.15966796875,
+          "probability_max_abs_error": 0.0005516764917031236,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 3,
+          "public_result_equal": false
+        },
+        {
+          "case": "twenty_options",
+          "questions": 1,
+          "sequence_length": 137,
+          "input_sha256": "f20fe7490e99ad43d4d1c5a9657f095968b904c9a60e8b3e02fce0544126f838",
+          "tokens_equal": true,
+          "logits_max_abs_error": 0.005983829498291016,
+          "action_logits_max_abs_error": 1.384765625,
+          "probability_max_abs_error": 1.541733241969172e-31,
+          "action_probability_max_abs_error": 0.0,
+          "argmax_agreements": 1,
+          "public_result_equal": true
+        }
+      ],
+      "stability": {
+        "calls": 100,
+        "elapsed_seconds": 3.521159500000067,
+        "finite": true,
+        "deterministic": true,
+        "active_bytes_before": 842602138,
+        "active_bytes_after": 842602138,
+        "active_growth_bytes": 0,
+        "peak_bytes": 1648506739
+      }
+    }
+  ]
+}