openthai-systemone 0.1.0__tar.gz

openthai_systemone-0.1.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 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
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

openthai_systemone-0.1.0/PKG-INFO

@@ -0,0 +1,104 @@
+Metadata-Version: 2.4
+Name: openthai-systemone
+Version: 0.1.0
+Summary: OpenThai-SystemOne: an open Thai/English System-One decision model (256-way slot softmax on Qwen3.5-0.8B)
+Author: iApp Technology / OpenThai
+License: Apache-2.0
+Project-URL: Homepage, https://iapp.co.th/docs/llm/openthai-systemone
+Project-URL: Repository, https://github.com/iapp-technology/openthai-systemone
+Project-URL: Model, https://huggingface.co/iapp/OpenThai-SystemOne
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: torch>=2.4
+Requires-Dist: transformers>=5.0
+Requires-Dist: numpy
+Requires-Dist: pydantic>=2
+Provides-Extra: server
+Requires-Dist: fastapi; extra == "server"
+Requires-Dist: uvicorn; extra == "server"
+Provides-Extra: train
+Requires-Dist: datasets; extra == "train"
+Requires-Dist: accelerate; extra == "train"
+Requires-Dist: pyyaml; extra == "train"
+Requires-Dist: tqdm; extra == "train"
+Requires-Dist: openai; extra == "train"
+Provides-Extra: dev
+Requires-Dist: pytest; extra == "dev"
+Dynamic: license-file
+
+# OpenThai-SystemOne
+
+![OpenThai-SystemOne launch: Jev ภาษาไทยมาแล้ว — open Thai + English System One decision model](https://raw.githubusercontent.com/iapp-technology/openthai-systemone/main/docs/assets/fb_launch_v3_1x1.png)
+
+Open-source Thai + English **System One decision model**: a Qwen3.5-0.8B text tower whose LM head is replaced by a
+**256-way slot softmax**, continued-pretrained on Thai, and trained to answer typed questions (`choice` / `score` /
+`noul`) about an arbitrary *state* in a single forward pass with calibrated probabilities. The request/response contract
+mirrors TypeSafe AI's `POST /v1/systemone` (Jev), so SDK code written for TypeSafe can point at this model.
+
+## Quickstart
+
+```bash
+pip install "git+https://github.com/iapp-technology/openthai-systemone"      # or: pip install openthai-systemone
+```
+
+```python
+from openthai_systemone import SystemOneClient, Choice, Score, Noul
+
+client = SystemOneClient("iapp/OpenThai-SystemOne")          # CUDA, MPS or CPU
+resp = client.system_one(
+    state={"ticket": "โดนหักเงินซ้ำสองครั้ง ขอเงินคืนด่วน โทรไปสามรอบแล้วไม่มีใครรับ"},
+    questions={
+        "department": Choice(instructions="ทีมใดควรรับผิดชอบ", criteria={"billing": "การเงิน/คืนเงิน", "technical": "ระบบใช้งานไม่ได้", "sales": None}),
+        "frustration": Score(instructions="ลูกค้าหงุดหงิดแค่ไหน", criteria=["ใจเย็น", "หงุดหงิดแต่สุภาพ", "โกรธมาก"]),
+        "refund": Noul(instructions="ลูกค้าขอเงินคืนอย่างชัดเจนหรือไม่"),
+    },
+)
+print(resp.answers["department"].choice, resp.answers["department"].probabilities)
+print(resp.answers["frustration"].score, resp.answers["refund"].noul)
+```
+
+HTTP server with the same JSON contract as TypeSafe's `POST /v1/systemone`:
+
+```bash
+OPENTHAI_SYSTEMONE_MODEL=iapp/OpenThai-SystemOne uvicorn openthai_systemone.server:app --port 8000
+```
+
+Model card with benchmark tables: https://huggingface.co/iapp/OpenThai-SystemOne. Design rationale: `docs/PLAN.md`.
+
+## Layout
+
+```
+openthai_systemone/     pip package: types (Choice/Score/Noul), formatting (prompt + slots), modeling (SlotHead), client, server
+scripts/00..07_*.py    one script per pipeline stage (see below)
+configs/*.yaml         CPT / SFT / calibration hyper-parameters (1x H100)
+tests/                 CPU unit tests
+```
+
+## Pipeline
+
+| stage | script | where | output |
+|---|---|---|---|
+| 0 surgery | `scripts/00_prepare_base.py` | CPU | `base/qwen3.5-0.8b-text` (vision encoder dropped) |
+| 1 CPT data | `scripts/01_cpt_data.py` | CPU, needs internet | `data/cpt/shard-*.bin` (5B Thai-heavy tokens) |
+| 1 CPT | `scripts/02_cpt_train.py --config configs/cpt.yaml` | 1x H100, ~1 day | `runs/cpt/latest` |
+| 2 decision data | `scripts/03_decision_data.py`, `scripts/03b_synth_generate.py` | CPU + LLM endpoint | `data/decision/*.jsonl`, `data/synth/*.jsonl` |
+| 2 SFT | `scripts/04_decision_train.py --config configs/sft.yaml` | 1x H100, ~0.5 day | `runs/sft/latest` |
+| 3 calibration | `scripts/04_decision_train.py --config configs/calib.yaml --init runs/sft/latest` | 1x H100, hours | `runs/calib/latest` |
+| 4 eval | `scripts/06_eval.py --model runs/calib/latest` | GPU | `runs/eval.json` |
+| 5 release | `scripts/07_export_release.py --ckpt runs/calib/latest --out release/... --push iapp/OpenThai-SystemOne` | – | HF Hub |
+
+## Dev
+
+```bash
+uv venv -p 3.12 .venv && source .venv/bin/activate
+uv pip install -e ".[server,train,dev]"
+pytest -q
+```
+
+Synthetic data needs an OpenAI-compatible endpoint:
+
+```bash
+export SYNTH_BASE_URL=http://localhost:11436/v1 SYNTH_MODEL=openthai2.0-qwen3.8-27b SYNTH_API_KEY=...
+python scripts/03b_synth_generate.py --out data/synth --n 600000 --concurrency 32
+```

openthai_systemone-0.1.0/README.md

@@ -0,0 +1,75 @@
+# OpenThai-SystemOne
+
+![OpenThai-SystemOne launch: Jev ภาษาไทยมาแล้ว — open Thai + English System One decision model](https://raw.githubusercontent.com/iapp-technology/openthai-systemone/main/docs/assets/fb_launch_v3_1x1.png)
+
+Open-source Thai + English **System One decision model**: a Qwen3.5-0.8B text tower whose LM head is replaced by a
+**256-way slot softmax**, continued-pretrained on Thai, and trained to answer typed questions (`choice` / `score` /
+`noul`) about an arbitrary *state* in a single forward pass with calibrated probabilities. The request/response contract
+mirrors TypeSafe AI's `POST /v1/systemone` (Jev), so SDK code written for TypeSafe can point at this model.
+
+## Quickstart
+
+```bash
+pip install "git+https://github.com/iapp-technology/openthai-systemone"      # or: pip install openthai-systemone
+```
+
+```python
+from openthai_systemone import SystemOneClient, Choice, Score, Noul
+
+client = SystemOneClient("iapp/OpenThai-SystemOne")          # CUDA, MPS or CPU
+resp = client.system_one(
+    state={"ticket": "โดนหักเงินซ้ำสองครั้ง ขอเงินคืนด่วน โทรไปสามรอบแล้วไม่มีใครรับ"},
+    questions={
+        "department": Choice(instructions="ทีมใดควรรับผิดชอบ", criteria={"billing": "การเงิน/คืนเงิน", "technical": "ระบบใช้งานไม่ได้", "sales": None}),
+        "frustration": Score(instructions="ลูกค้าหงุดหงิดแค่ไหน", criteria=["ใจเย็น", "หงุดหงิดแต่สุภาพ", "โกรธมาก"]),
+        "refund": Noul(instructions="ลูกค้าขอเงินคืนอย่างชัดเจนหรือไม่"),
+    },
+)
+print(resp.answers["department"].choice, resp.answers["department"].probabilities)
+print(resp.answers["frustration"].score, resp.answers["refund"].noul)
+```
+
+HTTP server with the same JSON contract as TypeSafe's `POST /v1/systemone`:
+
+```bash
+OPENTHAI_SYSTEMONE_MODEL=iapp/OpenThai-SystemOne uvicorn openthai_systemone.server:app --port 8000
+```
+
+Model card with benchmark tables: https://huggingface.co/iapp/OpenThai-SystemOne. Design rationale: `docs/PLAN.md`.
+
+## Layout
+
+```
+openthai_systemone/     pip package: types (Choice/Score/Noul), formatting (prompt + slots), modeling (SlotHead), client, server
+scripts/00..07_*.py    one script per pipeline stage (see below)
+configs/*.yaml         CPT / SFT / calibration hyper-parameters (1x H100)
+tests/                 CPU unit tests
+```
+
+## Pipeline
+
+| stage | script | where | output |
+|---|---|---|---|
+| 0 surgery | `scripts/00_prepare_base.py` | CPU | `base/qwen3.5-0.8b-text` (vision encoder dropped) |
+| 1 CPT data | `scripts/01_cpt_data.py` | CPU, needs internet | `data/cpt/shard-*.bin` (5B Thai-heavy tokens) |
+| 1 CPT | `scripts/02_cpt_train.py --config configs/cpt.yaml` | 1x H100, ~1 day | `runs/cpt/latest` |
+| 2 decision data | `scripts/03_decision_data.py`, `scripts/03b_synth_generate.py` | CPU + LLM endpoint | `data/decision/*.jsonl`, `data/synth/*.jsonl` |
+| 2 SFT | `scripts/04_decision_train.py --config configs/sft.yaml` | 1x H100, ~0.5 day | `runs/sft/latest` |
+| 3 calibration | `scripts/04_decision_train.py --config configs/calib.yaml --init runs/sft/latest` | 1x H100, hours | `runs/calib/latest` |
+| 4 eval | `scripts/06_eval.py --model runs/calib/latest` | GPU | `runs/eval.json` |
+| 5 release | `scripts/07_export_release.py --ckpt runs/calib/latest --out release/... --push iapp/OpenThai-SystemOne` | – | HF Hub |
+
+## Dev
+
+```bash
+uv venv -p 3.12 .venv && source .venv/bin/activate
+uv pip install -e ".[server,train,dev]"
+pytest -q
+```
+
+Synthetic data needs an OpenAI-compatible endpoint:
+
+```bash
+export SYNTH_BASE_URL=http://localhost:11436/v1 SYNTH_MODEL=openthai2.0-qwen3.8-27b SYNTH_API_KEY=...
+python scripts/03b_synth_generate.py --out data/synth --n 600000 --concurrency 32
+```

openthai_systemone-0.1.0/openthai_systemone/__init__.py

@@ -0,0 +1,21 @@
+from .types import Choice, Noul, Score, ChoiceAnswer, NoulAnswer, ScoreAnswer, SystemOneRequest, SystemOneResponse
+from .configuration import OpenThaiSystemOneConfig
+from .modeling import OpenThaiSystemOneForDecision
+from .formatting import Formatter, SPECIAL_TOKENS, N_SLOTS, ABSTAIN_SLOT
+from .client import SystemOneClient
+
+from transformers import AutoConfig, AutoModel
+
+try:  # make AutoConfig/AutoModel aware of the architecture when the package is installed
+    AutoConfig.register("openthai_systemone", OpenThaiSystemOneConfig)
+    AutoModel.register(OpenThaiSystemOneConfig, OpenThaiSystemOneForDecision)
+except ValueError:  # already registered (e.g. remote code + package both imported)
+    pass
+
+__version__ = "0.1.0"
+__all__ = [
+    "Choice", "Noul", "Score", "ChoiceAnswer", "NoulAnswer", "ScoreAnswer",
+    "SystemOneRequest", "SystemOneResponse",
+    "OpenThaiSystemOneConfig", "OpenThaiSystemOneForDecision",
+    "Formatter", "SPECIAL_TOKENS", "N_SLOTS", "ABSTAIN_SLOT", "SystemOneClient",
+]

openthai_systemone-0.1.0/openthai_systemone/client.py

@@ -0,0 +1,109 @@
+"""Local inference: `client.system_one(state, questions)` -> answers, mirroring the TypeSafe SDK."""
+from __future__ import annotations
+
+from typing import Any, Dict, List, Optional, Sequence, Union
+
+import torch
+from transformers import AutoTokenizer
+
+from .formatting import Encoded, Formatter, collate
+from .modeling import OpenThaiSystemOneForDecision, QTYPE_INDEX, confidence_from_probs, use_reference_kernels
+from .types import (
+    Answer,
+    Choice,
+    ChoiceAnswer,
+    Noul,
+    NoulAnswer,
+    Question,
+    Score,
+    ScoreAnswer,
+    SystemOneResponse,
+    Usage,
+    parse_question,
+)
+
+
+class SystemOneClient:
+    def __init__(
+        self,
+        model_path: str,
+        *,
+        device: Optional[str] = None,
+        dtype: Optional[torch.dtype] = None,
+        max_total_tokens: int = 65536,
+        max_state_tokens: int = 32768,
+        model_name: str = "openthai-systemone",
+    ):
+        if device is None:
+            device = "cuda" if torch.cuda.is_available() else ("mps" if torch.backends.mps.is_available() else "cpu")
+        if dtype is None:
+            dtype = torch.bfloat16 if device == "cuda" else torch.float32
+        self.device = device
+        if device != "cuda":
+            use_reference_kernels()
+        self.tok = AutoTokenizer.from_pretrained(model_path)
+        self.model = OpenThaiSystemOneForDecision.from_pretrained(model_path, dtype=dtype).to(device).eval()
+        self.fmt = Formatter(self.tok, max_total_tokens=max_total_tokens, max_state_tokens=max_state_tokens)
+        self.model_name = model_name
+        pad = self.tok.pad_token_id
+        self.pad_id = pad if pad is not None else (self.tok.eos_token_id or 0)
+
+    # ------------------------------------------------------------------ public API
+    @torch.no_grad()
+    def system_one(
+        self,
+        state: Union[str, Dict[str, Any], List[Any]],
+        questions: Dict[str, Union[Question, Dict[str, Any]]],
+    ) -> SystemOneResponse:
+        return self.system_one_batch([(state, questions)])[0]
+
+    @torch.no_grad()
+    def system_one_batch(
+        self,
+        items: Sequence[tuple],
+    ) -> List[SystemOneResponse]:
+        parsed = [(s, {k: parse_question(v) for k, v in qs.items()}) for s, qs in items]
+        encs = [self.fmt.encode(s, qs) for s, qs in parsed]
+        batch = collate(encs, self.pad_id)
+        qtypes = torch.full_like(batch["option_counts"], -1)
+        for b, e in enumerate(encs):
+            for qi, spec in enumerate(e.specs):
+                qtypes[b, qi] = QTYPE_INDEX[spec.qtype]
+        batch = {k: v.to(self.device) for k, v in batch.items()}
+        out = self.model(
+            input_ids=batch["input_ids"],
+            attention_mask=batch["attention_mask"],
+            answer_positions=batch["answer_positions"],
+            option_counts=batch["option_counts"],
+            qtypes=qtypes.to(self.device),
+            include_abstain=True,
+        )
+        probs = out.probs.float().cpu()
+        return [self._decode(e, probs[b]) for b, e in enumerate(encs)]
+
+    # ------------------------------------------------------------------ decoding
+    def _decode(self, enc: Encoded, probs: torch.Tensor) -> SystemOneResponse:
+        answers: Dict[str, Answer] = {}
+        for qi, spec in enumerate(enc.specs):
+            p = probs[qi]
+            k = len(spec.option_names)
+            abstain = float(p[self.model.config.abstain_slot])
+            pk = p[:k]
+            pk = pk / pk.sum().clamp(min=1e-12)  # renormalise over the real options
+            conf = confidence_from_probs(pk, k)
+            if spec.qtype == "noul":
+                answers[spec.qid] = NoulAnswer(noul=float(pk[1]))
+            elif spec.qtype == "choice":
+                prob_map = {spec.option_names[i]: float(pk[i]) for i in range(k)}
+                best = max(prob_map, key=prob_map.get)
+                answers[spec.qid] = ChoiceAnswer(choice=best, probabilities=prob_map, confidence=conf, abstain=abstain)
+            else:  # score
+                levels = torch.arange(k, dtype=torch.float32)
+                score = float((pk * levels).sum())
+                answers[spec.qid] = ScoreAnswer(
+                    score=score,
+                    legend={i: spec.option_descs[i] or str(i) for i in range(k)},
+                    probabilities={str(i): float(pk[i]) for i in range(k)},
+                    confidence=conf,
+                )
+        return SystemOneResponse(model=self.model_name, answers=answers, usage=Usage(input_tokens=enc.n_tokens))

openthai_systemone-0.1.0/openthai_systemone/configuration.py

@@ -0,0 +1,40 @@
+from __future__ import annotations
+
+from typing import Any, Dict, Optional
+
+from transformers import AutoConfig, PretrainedConfig
+
+
+class OpenThaiSystemOneConfig(PretrainedConfig):
+    """Config = a text-tower config (Qwen3.5 text by default) + slot-head settings."""
+
+    model_type = "openthai_systemone"
+    sub_configs = {"text_config": AutoConfig}
+
+    def __init__(
+        self,
+        text_config: Optional[Dict[str, Any] | PretrainedConfig] = None,
+        n_slots: int = 256,
+        abstain_slot: int = 255,
+        answer_token_id: Optional[int] = None,
+        head_bias: bool = True,
+        n_temperatures: int = 3,  # per question type: choice / score / noul
+        **kwargs,
+    ):
+        if isinstance(text_config, dict):
+            text_config = AutoConfig.for_model(**text_config) if "model_type" in text_config else AutoConfig.for_model("qwen3_5_text", **text_config)
+        self.text_config = text_config
+        self.n_slots = n_slots
+        self.abstain_slot = abstain_slot
+        self.answer_token_id = answer_token_id
+        self.head_bias = head_bias
+        self.n_temperatures = n_temperatures
+        super().__init__(**kwargs)
+
+    @property
+    def hidden_size(self) -> int:
+        return self.text_config.hidden_size
+
+    @property
+    def vocab_size(self) -> int:
+        return self.text_config.vocab_size

openthai_systemone-0.1.0/openthai_systemone/data/records.py

@@ -0,0 +1,95 @@
+"""Unified decision record format shared by all data builders and the SFT trainer.
+
+    {
+      "id": str, "source": str, "lang": "th"|"en"|"mixed"|..., "split": "train"|"eval",
+      "state": str | dict | list,
+      "questions": {qid: {"type": "choice"|"score"|"noul", "instructions": str, "criteria": ...}},
+      "labels":    {qid: option-name | level-index | bool}
+    }
+"""
+from __future__ import annotations
+
+import json
+import random
+from dataclasses import dataclass, field
+from typing import Any, Dict, Iterable, Iterator, List, Optional, Union
+
+# Instruction paraphrases: the model must not overfit to one phrasing per dataset.
+TH_CHOICE_TEMPLATES = [
+    "ข้อความนี้จัดอยู่ในหมวด{what}ใด",
+    "เลือก{what}ที่ตรงกับข้อความมากที่สุด",
+    "จาก state ด้านบน {what}ที่ถูกต้องคือข้อใด",
+    "ระบุ{what}ของข้อความนี้",
+]
+EN_CHOICE_TEMPLATES = [
+    "Which {what} best matches this text?",
+    "Classify the {what} of the state above.",
+    "Select the correct {what}.",
+    "What is the {what}?",
+]
+TH_NOUL_TEMPLATES = ["ข้อความนี้{what}หรือไม่", "จริงหรือไม่ว่า ข้อความนี้{what}", "state นี้{what}ใช่หรือไม่"]
+EN_NOUL_TEMPLATES = ["Is this text {what}?", "Does the state {what}?", "True or false: the text {what}."]
+
+
+@dataclass
+class Record:
+    id: str
+    source: str
+    lang: str
+    state: Union[str, Dict[str, Any], List[Any]]
+    questions: Dict[str, Dict[str, Any]]
+    labels: Dict[str, Union[str, int, bool]]
+    split: str = "train"
+    meta: Dict[str, Any] = field(default_factory=dict)
+
+    def to_json(self) -> str:
+        return json.dumps(self.__dict__, ensure_ascii=False)
+
+    @staticmethod
+    def from_json(line: str) -> "Record":
+        d = json.loads(line)
+        known = {"id", "source", "lang", "state", "questions", "labels", "split", "meta"}
+        meta = dict(d.get("meta", {}))
+        meta.update({k: v for k, v in d.items() if k not in known})
+        return Record(**{k: d[k] for k in ("id", "source", "lang", "state", "questions", "labels")},
+                      split=d.get("split", "train"), meta=meta)
+
+
+def choice_q(instructions: str, options: Iterable[str], descs: Optional[Dict[str, Optional[str]]] = None) -> Dict[str, Any]:
+    descs = descs or {}
+    return {"type": "choice", "instructions": instructions, "criteria": {o: descs.get(o) for o in options}}
+
+
+def score_q(instructions: str, levels: List[str]) -> Dict[str, Any]:
+    return {"type": "score", "instructions": instructions, "criteria": list(levels)}
+
+
+def noul_q(instructions: str, true_desc: Optional[str] = None, false_desc: Optional[str] = None) -> Dict[str, Any]:
+    q: Dict[str, Any] = {"type": "noul", "instructions": instructions}
+    if true_desc or false_desc:
+        q["criteria"] = {"true": true_desc, "false": false_desc}
+    return q
+
+
+def pick_template(rng: random.Random, lang: str, kind: str, what: str) -> str:
+    if kind == "choice":
+        t = TH_CHOICE_TEMPLATES if lang == "th" else EN_CHOICE_TEMPLATES
+    else:
+        t = TH_NOUL_TEMPLATES if lang == "th" else EN_NOUL_TEMPLATES
+    return rng.choice(t).format(what=what)
+
+
+def write_jsonl(path, records: Iterable[Record]) -> int:
+    n = 0
+    with open(path, "w", encoding="utf-8") as f:
+        for r in records:
+            f.write(r.to_json() + "\n")
+            n += 1
+    return n
+
+
+def read_jsonl(path) -> Iterator[Record]:
+    with open(path, encoding="utf-8") as f:
+        for line in f:
+            if line.strip():
+                yield Record.from_json(line)