Release RouteFM 1.0 model artifacts

LICENSE

@@ -0,0 +1,201 @@
+                                 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.

NOTICE

@@ -0,0 +1,4 @@
+RouteFM
+Copyright 2026 RouteFM contributors
+
+This product is licensed under the Apache License, Version 2.0.

README.md

@@ -1,3 +1,123 @@
 ---
 license: apache-2.0
+library_name: routefm
+tags:
+  - model-routing
+  - llm-routing
+  - multimodal-routing
+  - pytorch
+  - safetensors
 ---
+
+# RouteFM 1.0
+
+RouteFM is a pretrained in-context model router. Given a pool of candidate
+models, observations of their quality and cost on earlier Context queries, and
+the embedding of a new Target query, RouteFM predicts each candidate's quality
+and relative cost. The default decision rule selects the highest predicted
+quality.
+
+This repository is the canonical distribution point for the frozen RouteFM
+1.0 weights. Source code, data schemas, training configurations, benchmark
+splits, and evaluation scripts are maintained at
+[LAMDA-Model-Reuse/RouteFM](https://github.com/LAMDA-Model-Reuse/RouteFM).
+
+## Released variants
+
+| Variant | Compatible query encoder | Dimension | Query modality | Parameters |
+| --- | --- | ---: | --- | ---: |
+| Qwen RouteFM | Qwen3-VL-Embedding-8B | 4096 | Joint text and image | 12,781,059 |
+| BGE RouteFM | BAAI/bge-base-en-v1.5, CLS pooling | 768 | Text | 11,070,467 |
+
+The query encoders are frozen external feature extractors. They are not
+included here, and RouteFM is not a fine-tune of either encoder. Users must
+obtain the encoders or a compatible embedding service separately and follow
+their licenses and access terms.
+
+## Files and integrity
+
+Each variant contains a canonical `model.safetensors` and `config.json`.
+`legacy/` contains the exact PyTorch checkpoint bytes distributed in the
+GitHub v1.0.0 release. `manifest.json` records sizes and SHA-256 digests for
+all artifacts. The safetensors files contain the same model tensors as their
+legacy checkpoints; optimizer state is not distributed.
+
+The RouteFM package downloads the safetensors artifact from an immutable Hub
+revision, validates its SHA-256 digest, and caches it through
+`huggingface_hub`:
+
+```bash
+python -m pip install git+https://github.com/LAMDA-Model-Reuse/RouteFM.git
+
+routefm-predict --encoder qwen --input my_episode.npz \
+  --output predictions.json --device cpu
+```
+
+The complete custom-data schema and embedding instructions are documented in
+[`docs/CUSTOM_DATA.md`](https://github.com/LAMDA-Model-Reuse/RouteFM/blob/v1.0.0/docs/CUSTOM_DATA.md).
+
+## Training
+
+Both variants were trained from random initialization in one continuous
+10,000-update run. All router modules were trained jointly. The external query
+encoders were frozen. The configured pretraining mixture is LLMRouterBench
+45%, RouterBench 25%, RouterEval 25%, and MixInstruct 5%.
+
+The exact configurations and preprocessing contract are available in
+[`docs/PRETRAINING.md`](https://github.com/LAMDA-Model-Reuse/RouteFM/blob/v1.0.0/docs/PRETRAINING.md).
+The underlying third-party datasets and embedding models are not
+redistributed by this repository.
+
+## MMR-Bench V1 evaluation
+
+The following results use 6,223 Target queries and the published seed-31010
+within-dataset 40% Context / 60% Target split. Quality is target-count weighted
+across benchmark datasets.
+
+| Context protocol | Qwen RouteFM | BGE RouteFM | Context-mean baseline |
+| --- | ---: | ---: | ---: |
+| K=8 | 0.7463 | 0.7413 | 0.7156 |
+| K=16 | 0.7527 | 0.7422 | 0.7458 |
+| K=32 | 0.7546 | 0.7469 | 0.7496 |
+| K=64 | 0.7585 | 0.7487 | 0.7482 |
+| Complete 40% Context pool | 0.7614 | 0.7492 | 0.7577 |
+
+These results are retrospective. MMR-Bench was inspected during the broader
+research process, and seed 31010 was selected after a ten-seed split
+sensitivity sweep. They must be described as a post-selected illustrative
+split, not an unbiased multi-seed estimate or untouched holdout. No MMR-Bench
+score or cost cell enters the included pretraining recipe. The complete
+protocol and exact IDs are in
+[`docs/MMRBENCH_V1.md`](https://github.com/LAMDA-Model-Reuse/RouteFM/blob/v1.0.0/docs/MMRBENCH_V1.md).
+
+## Intended use and limitations
+
+RouteFM is intended for research on routing among a user-supplied candidate
+pool with observed Context outcomes. It does not call candidate models,
+generate query embeddings, or establish that a candidate is safe or suitable.
+
+- Candidate order and embedding family must remain consistent within an episode.
+- Each candidate needs at least one valid Context observation.
+- Predicted cost is relative and is not a calibrated monetary or latency estimate.
+- The default decision rule ignores predicted cost and chooses maximum predicted quality.
+- Performance may change with encoder revisions, embedding services, domains,
+  languages, candidate pools, and Context sizes not represented during training.
+- This release does not include a mixture-of-experts router.
+
+## License and citation
+
+RouteFM source code and released routing weights are licensed under Apache-2.0.
+Third-party assets retain their own licenses and terms. See
+[`THIRD_PARTY_NOTICES.md`](https://github.com/LAMDA-Model-Reuse/RouteFM/blob/v1.0.0/THIRD_PARTY_NOTICES.md).
+
+Until a paper citation is published, cite the versioned software artifact:
+
+```bibtex
+@software{routefm2026,
+  title  = {RouteFM: Pretrained In-Context Model Routing},
+  author = {{RouteFM Contributors}},
+  year   = {2026},
+  url    = {https://github.com/LAMDA-Model-Reuse/RouteFM},
+  version = {1.0.0}
+}
+```

THIRD_PARTY_NOTICES.md

@@ -0,0 +1,19 @@
+# Third-Party Notices
+
+RouteFM is distributed under the Apache License, Version 2.0. This license
+applies to the RouteFM source code and the RouteFM routing weights included in
+this repository. It does not replace or modify the licenses, terms of use, or
+access requirements of third-party assets.
+
+The repository supports or documents interoperability with the following
+external assets, which are not redistributed here:
+
+- Qwen3-VL-Embedding-8B and compatible embedding services;
+- BAAI/bge-base-en-v1.5;
+- MMR-Bench V1, including its records and images;
+- LLMRouterBench, RouterBench, RouterEval, and MixInstruct pretraining data;
+- Python dependencies declared in `pyproject.toml`.
+
+Users must obtain these assets from their respective providers and comply with
+the providers' applicable licenses and terms. Names and trademarks are used
+only to identify compatible models, datasets, benchmarks, and software.

bge/config.json

@@ -0,0 +1,41 @@
+{
+  "encoder": "bge",
+  "format": "routefm_safetensors",
+  "model_config": {
+    "architecture": "profile_context",
+    "auxiliary_distribution_head": false,
+    "auxiliary_extended_observation_encoder": false,
+    "capability_tokens": 24,
+    "context_scale_conditioning": false,
+    "dropout": 0.1,
+    "ffn_dim": 1024,
+    "heads": 8,
+    "hidden_dim": 256,
+    "local_context_gate_hidden": 64,
+    "local_context_gate_init": 0.25,
+    "local_context_layers": 2,
+    "local_context_mode": "parallel_static",
+    "local_context_residual_init": 0.1,
+    "local_context_semantic_bias": 0.0,
+    "local_context_topk": 0,
+    "observation_features": 2,
+    "observation_schema": "score_cost",
+    "outcome_residual_init": 3.0,
+    "pool_layers": 3,
+    "profile_layers": 4,
+    "projection_dim": 512,
+    "quality_bins": 21,
+    "query_dim": 768,
+    "readout_layers": 2,
+    "target_query_scale": 1.0,
+    "use_pool_transformer": true,
+    "utility_cost_weight": 0.0,
+    "utility_uncertainty_weight": 0.0
+  },
+  "model_version": "1.0.0",
+  "parameter_count": 11070467,
+  "release_version": "RouteFM 1.0",
+  "source_checkpoint": "legacy/routefm_bge.pt",
+  "source_checkpoint_sha256": "0f7ce1b269499a6938a5722511a489eb1a701e17dfd2605f49d47a19edeffcde",
+  "step": 10000
+}

bge/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3245698f2f6bd4cb607cc3af8875d63ba04172f33abd25b286d1b2968c679662
+size 44302700

legacy/routefm_bge.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f7ce1b269499a6938a5722511a489eb1a701e17dfd2605f49d47a19edeffcde
+size 44355866

legacy/routefm_qwen.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e28d6203e99cf5f10f42293c561b80d17e866bc7bd69a696928ea836e6b7cccc
+size 51199466

manifest.json

@@ -0,0 +1,29 @@
+{
+  "artifacts": {
+    "bge": {
+      "config": "bge/config.json",
+      "config_sha256": "6cdc9e9d5ccad4ed9e2543e26ea5adf7def5830189a25e3349ae183c5e7535ea",
+      "legacy_checkpoint": "legacy/routefm_bge.pt",
+      "legacy_checkpoint_sha256": "0f7ce1b269499a6938a5722511a489eb1a701e17dfd2605f49d47a19edeffcde",
+      "legacy_checkpoint_size_bytes": 44355866,
+      "weights": "bge/model.safetensors",
+      "weights_sha256": "3245698f2f6bd4cb607cc3af8875d63ba04172f33abd25b286d1b2968c679662",
+      "weights_size_bytes": 44302700
+    },
+    "qwen": {
+      "config": "qwen/config.json",
+      "config_sha256": "0a0f63b954fa3e4472fdd8c51abff33bbc6a12bf0ea267eab0a4604ebcaf6231",
+      "legacy_checkpoint": "legacy/routefm_qwen.pt",
+      "legacy_checkpoint_sha256": "e28d6203e99cf5f10f42293c561b80d17e866bc7bd69a696928ea836e6b7cccc",
+      "legacy_checkpoint_size_bytes": 51199466,
+      "weights": "qwen/model.safetensors",
+      "weights_sha256": "881b648ead20565da38a7f6ee9516dd0a45d3747add41c44d300ac24433b9cb8",
+      "weights_size_bytes": 51145068
+    }
+  },
+  "code_repository": "https://github.com/LAMDA-Model-Reuse/RouteFM",
+  "license": "Apache-2.0",
+  "model_version": "1.0.0",
+  "name": "RouteFM",
+  "source_code_revision": "v1.0.0"
+}

qwen/config.json

@@ -0,0 +1,69 @@
+{
+  "encoder": "qwen",
+  "format": "routefm_safetensors",
+  "model_config": {
+    "architecture": "profile_context",
+    "auxiliary_distribution_head": false,
+    "auxiliary_extended_observation_encoder": false,
+    "bilinear_scale": 1.0,
+    "capability_tokens": 24,
+    "conditional_residual_init": 0.25,
+    "conditional_residual_max": 0.5,
+    "context_refinement_rounds": 0,
+    "context_refinement_shared": false,
+    "context_scale_conditioning": false,
+    "context_scale_max": 1024,
+    "context_scale_residual_init": 0.1,
+    "dropout": 0.1,
+    "ensemble_adapter_dim": 64,
+    "ensemble_branch_mode": "late_residual",
+    "ensemble_context_policies": [],
+    "ensemble_gate_hidden": 32,
+    "ensemble_gate_mode": "none",
+    "ensemble_member_weights": [],
+    "ensemble_members": 1,
+    "ensemble_residual_init": 0.001,
+    "ensemble_semantic_topk": 16,
+    "ffn_dim": 1024,
+    "grid_layers": 2,
+    "grid_residual_init": 0.1,
+    "heads": 8,
+    "hidden_dim": 256,
+    "kernel_scale": 4.0,
+    "kernel_temperature": 0.1,
+    "local_context_gate_hidden": 64,
+    "local_context_gate_init": 0.25,
+    "local_context_layers": 2,
+    "local_context_mode": "parallel_static",
+    "local_context_residual_init": 0.1,
+    "local_context_semantic_bias": 0.0,
+    "local_context_topk": 0,
+    "observation_features": 2,
+    "observation_schema": "score_cost",
+    "outcome_residual_init": 3.0,
+    "pool_layers": 3,
+    "profile_layers": 4,
+    "projection_dim": 512,
+    "quality_bins": 21,
+    "query_dim": 4096,
+    "readout_layers": 2,
+    "ridge_feature_dim": 16,
+    "ridge_lambda": 1.0,
+    "similarity_adapter_hidden": 32,
+    "similarity_logit_scale": 0.75,
+    "similarity_sigma": 0.12,
+    "similarity_temperature": 0.1,
+    "target_profile_feedback_residual_init": 0.1,
+    "target_profile_feedback_rounds": 0,
+    "target_query_scale": 1.0,
+    "use_pool_transformer": true,
+    "utility_cost_weight": 0.0,
+    "utility_uncertainty_weight": 0.0
+  },
+  "model_version": "1.0.0",
+  "parameter_count": 12781059,
+  "release_version": "RouteFM 1.0",
+  "source_checkpoint": "legacy/routefm_qwen.pt",
+  "source_checkpoint_sha256": "e28d6203e99cf5f10f42293c561b80d17e866bc7bd69a696928ea836e6b7cccc",
+  "step": 10000
+}

qwen/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:881b648ead20565da38a7f6ee9516dd0a45d3747add41c44d300ac24433b9cb8
+size 51145068