Laya-Android v0.1.0 model card (no weights yet)

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 +1,29 @@
+Laya-Android
+Copyright 2026 Jaeyeon Byun
+
+This project is licensed under the Apache License, Version 2.0 (see LICENSE).
+
+This project builds on the following third-party work:
+
+1. Laya (code and model weights) by Convai Innovations
+   https://github.com/NandhaKishorM/laya
+   https://huggingface.co/convaiinnovations/laya
+   Licensed under the Apache License, Version 2.0.
+   Laya-Android fine-tunes the `multilingual` checkpoint of convaiinnovations/laya
+   (revision 7b928d828b7b0e022f929d9bd2e44165aa270148) and uses the `laya` Python
+   package (0.3.28) for model code, input encoding and the training loop.
+   Laya-Android is not affiliated with or endorsed by Convai Innovations.
+
+2. mmBERT-base (encoder backbone inside the Laya multilingual checkpoint)
+   https://huggingface.co/jhu-clsp/mmBERT-base
+   Licensed under the MIT License.
+
+3. AndroidControl dataset by Google Research
+   https://github.com/google-research/google-research/tree/master/android_control
+   Copyright 2024 The Google Research Authors.
+   Licensed under the Apache License, Version 2.0.
+   Li et al., "On the Effects of Data Scale on Computer Control Agents", arXiv:2406.03679.
+   The dataset is not redistributed in this repository.
+
+4. Accessibility proto field layout follows google-deepmind/android_env
+   (Apache License, Version 2.0). No android_env code is included.

README.md

@@ -0,0 +1,161 @@
+---
+license: apache-2.0
+base_model: convaiinnovations/laya
+language:
+  - en
+tags:
+  - android
+  - gui-agent
+  - accessibility
+  - decision-model
+  - laya
+  - androidcontrol
+---
+
+# Laya-Android
+
+> **Status: preliminary.** Training and evaluation are in progress. Final weights and benchmark results are not released yet. Numbers below are validation-only.
+
+## Model Description
+
+Laya-Android is a 322M non-generative decision model specialized for Android GUI action selection from accessibility-based interaction trajectories.
+
+Given a high-level goal, the current accessibility tree (as a list of actionable UI candidates) and recent action history, it predicts the next operation and, for taps, the target UI element. It does this in a single forward pass, without autoregressive text generation and without screenshots.
+
+It is a fine-tune of [Laya](https://huggingface.co/convaiinnovations/laya) by Convai Innovations. The architecture is Laya's, unchanged; Laya-Android contributes the Android data pipeline, candidate construction and the fine-tuned weights. It is not a vision model, a VLM or a world model.
+
+## Model Details
+
+| | |
+|---|---|
+| Developed by | Jaeyeon Byun (GitHub: Byun11) |
+| Model type | Non-generative typed-decision model (bidirectional encoder + option-marker decision head) |
+| Parameters | ~322M |
+| Encoder | mmBERT-base |
+| License | Apache-2.0 |
+| Code | https://github.com/Byun11/laya-android |
+| Sequence budget | `max_len` 3072, `head_max_len` 1536 |
+
+## Base Model
+
+- `convaiinnovations/laya`, subfolder `multilingual`
+- revision `7b928d828b7b0e022f929d9bd2e44165aa270148`
+- `laya` Python package 0.3.28
+- Laya is Apache-2.0; mmBERT-base is MIT. Laya-Android is not affiliated with or endorsed by Convai Innovations.
+
+## Intended Use
+
+- research on lightweight Android GUI action-selection policies
+- accessibility-based mobile agents
+- fast System-1 style decision models, for example as a step selector under a planner
+
+## Out-of-Scope Use
+
+- Operating real devices or accounts unsupervised, especially for payments, messaging, account or security settings, or other irreversible actions.
+- Tasks that need visual understanding of the screen (images, canvases, games, custom-drawn UI).
+- Generating free text: the model does not produce `INPUT_TEXT` content or `OPEN_APP` names.
+- Any claim of end-to-end task success: v0 is evaluated offline, step by step.
+
+## Training Data
+
+[AndroidControl](https://github.com/google-research/google-research/tree/master/android_control) (Google Research, Apache-2.0) only, with the official splits.
+
+- train: 13,594 episodes, 74,714 raw actions (9 zero-action episodes excluded), 88,308 decision steps including a synthetic `DONE` after each episode, 132,448 training items (operation and target questions).
+- Click grounding coverage on train: 94.76%. Candidates at groundable clicks: median 15, max 246; none truncated.
+- Screenshots and low-level step instructions are not used as model input.
+
+## Training Procedure
+
+- Supervised fine-tuning (behavior cloning), cross-entropy over options, hard labels.
+- bf16, gradient checkpointing, effective batch 32, encoder lr 2.5e-5, head lr 1e-4, cosine schedule.
+- Option order shuffled during training.
+- No online RL, no oversampling, no class weighting, no candidate truncation.
+- Up to 3 epochs; the checkpoint is selected by validation joint accuracy only, and temperatures are fit on validation only.
+- Hardware: 1× RTX 4090 24GB, about 55 minutes per epoch.
+
+## Input Format
+
+Each decision step is two Laya choice questions over one shared state:
+
+```json
+{
+  "state": {
+    "goal": "<high-level goal>",
+    "history": ["<last 3 actions>"],
+    "ui": ["[0] <label> (<class>) <flags>", "[1] ...", "..."]
+  },
+  "questions": {
+    "operation": {"type": "choice", "instructions": "Which operation should be performed next?",
+                  "criteria": {"CLICK": "tap a UI element", "...": "..."}},
+    "target": {"type": "choice", "instructions": "Which UI element should be targeted?",
+               "criteria": {"0": "<label> (<class>) <flags>", "1": "..."}}
+  }
+}
+```
+
+Use the repository's `src/candidates.py` and `src/serialization.py` to build exactly this input from an accessibility forest; other serializations are out of distribution.
+
+## Output Format
+
+A probability distribution over options for each question:
+
+- `operation`: one of `CLICK, LONG_PRESS, SCROLL_UP, SCROLL_DOWN, SCROLL_LEFT, SCROLL_RIGHT, OPEN_APP, INPUT_TEXT, BACK, HOME, WAIT, DONE`
+- `target`: the index of a UI candidate (asked for `CLICK` / `LONG_PRESS`)
+
+## Evaluation
+
+Offline, step-level, on AndroidControl, with the high-level goal only. Metrics: operation accuracy, operation macro-F1 with per-class precision/recall, target top-1 on groundable clicks, joint accuracy, grounding coverage, ECE and Brier, target top-1 by candidate count, and latency. Text payloads are not scored, so the numbers are not directly comparable to AndroidControl step accuracy reported elsewhere.
+
+### Preliminary (validation only)
+
+| AndroidControl validation (827 steps) | Base Laya (zero-shot) | Laya-Android, epoch 1 |
+|---|---:|---:|
+| joint accuracy | 0.098 | 0.605 |
+| target top-1 | 0.110 | 0.631 |
+| operation accuracy | 0.223 | 0.768 |
+| operation macro-F1 | 0.142 | 0.543 |
+| operation ECE | 0.604 | 0.027 |
+
+These are validation results, not held-out final test results. Test results will be added once the checkpoint is final.
+
+## Limitations
+
+- No visual input.
+- Relies on accessibility semantics; unlabeled icons are hard to distinguish.
+- Arbitrary text generation is unsupported.
+- Weak on actions with little supervision (for example `LONG_PRESS`, 157 training examples).
+- Target performance is bounded by candidate quality; about 5% of clicks cannot be grounded.
+- Custom-drawn or inaccessible UI may be unobservable.
+- v0 is AndroidControl-specialized and has not been evaluated online.
+
+## Ethical / Safety Considerations
+
+- A GUI action model can trigger real side effects on a device. Keep a human in the loop and gate sensitive or irreversible actions.
+- Accessibility trees can contain personal data shown on screen (messages, contacts, account details). Process them locally and do not log them unnecessarily.
+- Calibrated confidence (ECE above) can support abstention, but low confidence does not guarantee safety and high confidence does not guarantee correctness.
+- The training data comes from crowd-sourced demonstrations on a limited set of apps and may not reflect other apps, locales or accessibility needs.
+
+## Citation
+
+```bibtex
+@software{laya_android_2026,
+  title   = {Laya-Android},
+  author  = {Byun, Jaeyeon},
+  year    = {2026},
+  version = {0.1.0},
+  url     = {https://github.com/Byun11/laya-android}
+}
+
+@article{li2024effects,
+  title={On the Effects of Data Scale on Computer Control Agents},
+  author={Li, Wei and Bishop, William and Li, Alice and Rawles, Chris and Campbell-Ajala, Folawiyo and Tyamagundlu, Divya and Riva, Oriana},
+  journal={arXiv preprint arXiv:2406.03679},
+  year={2024}
+}
+```
+
+## Acknowledgements
+
+- Built on top of [Laya](https://github.com/NandhaKishorM/laya) by Convai Innovations (Apache-2.0).
+- Encoder: [mmBERT-base](https://huggingface.co/jhu-clsp/mmBERT-base) (MIT).
+- Data: [AndroidControl](https://github.com/google-research/google-research/tree/master/android_control) by Google Research (Apache-2.0).