Executor de avaliação, métricas e .gitignore para metadados de modelos

- harness/evaluate.py: partição congelada com SHA-256, amostra com seed,
  lote 1, truncamento comum, contagem de tokens, latência (sem aquecimento),
  pico de RAM e GPU; teste bloqueado até o congelamento da configuração.
- analysis/metrics.py: acurácia, macro F1, Brier, NLL, ECE, acurácia seletiva.
- Diagnósticos por resposta (metadata.diagnostics); Laya informa opções
  colapsadas pelo orçamento de cabeçalho; head_max_len via DMB_LAYA_HEAD_MAX_LEN.
- infra/ds/collect.py e gpu_queue --env para pilotos e rodadas.
- .gitignore: de models/ só metadados (cards, configs, licenças, índices);
  pesos, tokenizadores grandes, runtimes, fontes clonados e caches ignorados.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

models/rizzo_flow/rizzoaiacademy__rizzo-flow/55633c8cbd2b826bd3eefdeb05310450996649df/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 2026 XHToken
+
+   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.

models/rizzo_flow/rizzoaiacademy__rizzo-flow/55633c8cbd2b826bd3eefdeb05310450996649df/README.md

@@ -0,0 +1,166 @@
+---
+license: apache-2.0
+base_model:
+  - XHToken/Spark-X2.5-4B
+  - XHToken/Spark-X2.5-1.7B
+library_name: peft
+tags:
+  - gguf
+  - lora
+  - classification
+  - typed-decisions
+  - rizzo-flow
+language:
+  - en
+---
+
+# Rizzo Flow: Spark-X2.5 (4B and 1.7B) + typed-decisions LoRA
+
+These are LoRA fine-tunes of [XHToken/Spark-X2.5-4B](https://huggingface.co/XHToken/Spark-X2.5-4B)
+(recommended) and [XHToken/Spark-X2.5-1.7B](https://huggingface.co/XHToken/Spark-X2.5-1.7B), for [Rizzo Flow](https://github.com/Rizzo-AI-Academy/rizzo-flow). Rizzo Flow turns an unstructured
+state into typed decisions (`boolean`, `choice`, `score`) with probabilities, and it **generates no
+text**. The model runs one forward pass per question. The answer is the softmax over the logits of
+the answer letters (`A`, `B`, …) at the last prompt position.
+
+Training works the same way. There is no text target: the loss is a soft cross-entropy between the
+target distribution and the softmax over the answer letters. Only the LoRA adapters are trained.
+
+This model is not affiliated with TypeSafe, and it is not Jev.
+
+## Files
+
+| File | What it is |
+| --- | --- |
+| `model-0000*-of-00005.safetensors`, `config.json`, tokenizer files | Merged model, BF16 safetensors (8.2 GB): for the MLX backend and transformers. Same weights as the BF16 GGUF, bit for bit |
+| `spark-x2.5-4b-rizzo-flow-lora-q8_0.gguf` | Merged model, Q8_0 (4.4 GB): the default for `rizzo serve` |
+| `spark-x2.5-4b-rizzo-flow-lora-bf16.gguf` | Merged model, BF16 (8.2 GB) |
+| `spark-x2.5-4b-rizzo-flow-lora-q4_k_m.gguf` | Merged model, Q4_K_M (2.6 GB), for GPUs with little memory |
+| `lora/` | The PEFT adapter alone (r 16, alpha 32), to apply to the BF16 HF checkpoint |
+| `training/history.json` | Arguments, training loss and dev evaluations of the 4B run |
+| `spark-x2.5-1.7b-rizzo-flow-lora-q8_0.gguf` | 1.7B merged model, Q8_0 (1.8 GB) |
+| `spark-x2.5-1.7b-rizzo-flow-lora-bf16.gguf` | 1.7B merged model, BF16 (3.4 GB) |
+| `lora-1.7b/` | The 1.7B PEFT adapter alone (r 16, alpha 32) |
+| `training/history-1.7b.json` | The same for the 1.7B run |
+
+## Use
+
+These GGUF files are what [Rizzo Flow](https://github.com/Rizzo-AI-Academy/rizzo-flow) downloads
+by default. It pins this repository by commit and checks the sha256 of every file:
+
+```bash
+git clone https://github.com/Rizzo-AI-Academy/rizzo-flow && cd rizzo-flow
+uv sync --locked
+uv run rizzo download          # llama.cpp for this machine + this model, Q8_0 (4.4 GB)
+uv run rizzo serve             # → http://127.0.0.1:8017/playground
+```
+
+`--quant bf16` or `--quant q4_k_m` picks another file, `--size 1.7b` the small model (also in its own repository,
+[`rizzoaiacademy/rizzo-flow-1.7b`](https://huggingface.co/rizzoaiacademy/rizzo-flow-1.7b)), and
+`--weights base` the original Spark-X2.5 GGUF files. The server answers as `rizzo-flow-4b-q8_0`.
+
+**MLX** (Apple Silicon, or MLX-CUDA) uses the safetensors checkpoint at the root of this
+repository instead of the GGUF files:
+
+```bash
+uv sync --locked --extra mlx
+uv run --no-sync rizzo download --backend mlx      # the safetensors only, 8.2 GB
+uv run --no-sync rizzo serve --backend mlx --bits 8
+```
+
+It answers as `rizzo-flow-4b-bf16` (or `…-q8`/`…-q4` with `--bits`). On an RTX 5060 Ti with MLX-CUDA, BF16, the
+4B gives the same prompts (same hashes) and the same probabilities as llama.cpp BF16 within 0.001.
+
+**transformers** loads the checkpoint with the original Spark code (`trust_remote_code=True`,
+`modeling_spark.py` is XHToken's file, unchanged): `AutoModelForCausalLM.from_pretrained(
+"rizzoaiacademy/rizzo-flow", trust_remote_code=True, dtype="bfloat16")`. Use the prompts of
+Rizzo Flow and read the answer-letter logits at the last position; generating text is not what
+it was trained for.
+
+The GGUF files are plain GGUF (architecture `spark2_5`; Rizzo Flow runs them on llama.cpp release b11081), but the model is
+meant to be **read, not chatted with**: Rizzo Flow builds the prompt, then reads the logits of the
+answer letters. In a chat app it behaves like a lightly changed Spark-X2.5, and the fine-tune
+brings nothing there.
+
+The adapter was trained on the exact prompts of `PROMPT_VERSION = "spark-decisions-v3"`, commit
+`c30cc63` of rizzo-flow. Any other prompt format can make the adapter useless.
+
+## Results
+
+We ran the test split of the
+[`LocalLLaMA/typed-decisions`](https://huggingface.co/datasets/LocalLLaMA/typed-decisions) benchmark
+(config `all`, 400 cases, 2,000 decisions). The model is fine-tuned for typed decisions, so this
+is not a zero-shot result, but training used no data from these four workflows: the Open-Jev config
+that shares them (`workflow-controls-v1`) was excluded, and so was every training state containing
+text from the test set. Base and fine-tune were measured on the same machine: llama.cpp b11081,
+CUDA, RTX 5060 Ti.
+
+| | Accuracy ↑ | KL from gold ↓ | Brier ↓ | ECE ↓ | p50 per case |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| Spark-X2.5-4B Q8_0, base | 0.574 | 2.899 | 0.480 | 0.349 | 201 ms |
+| **This model, Q8_0** | **0.648** | **0.452** | **0.205** | **0.112** | 195 ms |
+| TypeSafe Jev 1.13.0 (from its dataset card, not re-measured) | 0.727 | 1.442 | 0.148 | – | – |
+
+Accuracy +0.074, 95% interval [+0.050, +0.101] (paired bootstrap over cases). On an RTX PRO 6000
+the same file scores 0.650 / 0.454 / 0.205 / 0.111: the difference between GPUs is below 0.002.
+
+**Q4_K_M** (made with `llama-quantize` b11081 from the BF16 GGUF, no importance matrix) scores
+0.650 / 0.436 / 0.201 / 0.093 on the RTX 5060 Ti, p50 198 ms: tied with Q8_0 on this benchmark
+(+0.002 [−0.011, +0.015]). On the original weights Q4_K_M cost 4 points on SemIf's fixtures, which
+we have not re-run here, so check it on your own data before relying on it.
+
+Per workflow (accuracy / KL): customer_service 0.710 / 0.350, invoice_processing 0.714 / 0.506,
+security_incidents 0.674 / 0.434, agent_trace_observability 0.500 / 0.526 (base: 0.366).
+By type: noul 0.738, choice 0.637, score 0.593.
+
+Dev split of the training mixture (600 questions): loss 1.928 → 0.393, accuracy 0.547 → 0.857.
+
+### 1.7B
+
+We measured the base model and the fine-tuned one on the same machine, with the same runtime
+(RTX PRO 6000; on the RTX 5060 Ti: 0.530 → 0.544, KL 3.031 → 0.694):
+
+| | Accuracy ↑ | KL from gold ↓ | Brier ↓ | ECE ↓ | p50 per case |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| Spark-X2.5-1.7B Q8_0, base | 0.531 | 3.032 | 0.494 | 0.348 | 50 ms |
+| **This model, 1.7B Q8_0** | **0.546** | **0.692** | **0.274** | **0.167** | 52 ms |
+
+On the 1.7B, fine-tuning mostly improves the probabilities (KL, Brier, ECE). Accuracy barely
+moves: +0.015 is within noise. The change is not uniform across workflows:
+agent_trace_observability goes 0.372 → 0.490 and customer_service 0.644 → 0.674, but
+security_incidents drops 0.608 → 0.520. By type, noul goes 0.603 → 0.675 and score 0.441 → 0.461,
+but choice drops 0.578 → 0.532. On the dev split, loss goes 3.117 → 0.460 and accuracy
+0.408 → 0.820. Use the 4B when accuracy matters.
+
+The probabilities are **not calibrated**. Fit temperature scaling on data from your own domain
+(`rizzo calibrate`) before you use thresholds. The accuracy is still below Jev's published
+number, and the gold labels of the benchmark come from a ~4B teacher model, not from humans. We
+have not yet re-run the smoke and SemIf fixtures on this model.
+
+## Training
+
+- Base: Spark-X2.5-4B BF16, rev. `0bcb356`. LoRA r 16, alpha 32, dropout 0.05, on all linear
+  layers of attention and MLP. Embedding and `lm_head` are not trained because they are tied.
+  32.4M trainable parameters.
+- AdamW, lr 5e-5, 3% warmup then cosine decay, no weight decay, clip 1.0. 16 questions per step.
+  1 epoch = 1,770 steps. Seed 0.
+- 28,321 training questions (14.7M prompt tokens, max 2,048 tokens per prompt), from
+  [`tasksource/procedural-typed-decisions`](https://huggingface.co/datasets/tasksource/procedural-typed-decisions)
+  (Apache-2.0), [`ZefanCai/Open-Jev`](https://huggingface.co/datasets/ZefanCai/Open-Jev)
+  `release-v2-redistributable` (CC0) and 12 configs of
+  [`Praveenrajus/jev-bench`](https://huggingface.co/datasets/Praveenrajus/jev-bench). None of the
+  labels come from Jev or TypeSafe. We removed every training state that contains text from the
+  evaluation sets.
+- Hardware: 1× RTX PRO 6000 Blackwell 96 GB, no gradient checkpointing. 4B: batch 4 / 2,048
+  tokens per forward, ~6,700 prompt tokens/s, 41 minutes, 38 GiB peak. 1.7B (rev. `14d6e83`,
+  16.7M trainable parameters, same data and hyperparameters): batch 8 / 3,072 tokens,
+  ~15,500 tokens/s, 19 minutes, 22 GiB peak.
+
+Full pipeline and reasoning:
+[`docs/training.md`](https://github.com/Rizzo-AI-Academy/rizzo-flow/blob/main/docs/training.md).
+
+## Licenses of the training data
+
+The base model is Apache-2.0. Among the jev-bench configs, `stsb` is **CC-BY-SA**.
+`measuring_hate_speech` and the HelpSteer2 configs are CC-BY-4.0: *Measuring Hate Speech*
+(Kennedy et al., 2020; Sachdeva et al., 2022) and *HelpSteer2* (Wang et al., 2024, NVIDIA).
+Check whether these terms matter for your use before you redistribute derivatives.