Release Vela on the unified Encoder base

LICENSE

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

README.md

@@ -1,13 +1,20 @@
 ---
+library_name: transformers
 license: apache-2.0
-library_name: pytorch
-pipeline_tag: text-ranking
-base_model: llm-semantic-router/mmbert-rerank-32k-2d-matryoshka
+pipeline_tag: text-classification
+base_model: llm-semantic-router/Vela-1.0-Encoder-307M
+base_model_relation: finetune
+language:
+- en
+- ar
+- es
+- ja
+- zh
 tags:
-  - semantic-router
-  - reranker
-  - multilingual
-  - matryoshka
+- modernbert
+- semantic-router
+- vela
+- matryoshka
 ---
 
 <div align="center">
@@ -20,55 +27,83 @@ tags:
   </p>
 </div>
 
-# Vela-1.0-Encoder-307M-Reranker
+# Vela Reranker
 
-**Bring the most relevant context to the front.**
+Vela Reranker scores query–passage pairs after an initial retrieval stage. Built from **Vela Base**, it offers 20 depth/dimension combinations. Higher raw scores rank a passage as more relevant to its query.
 
-307M · Multilingual · 32K context
-
-Vela Reranker scores candidate documents against a query so your application can put useful evidence first. Use it after retrieval to choose the context that reaches your language model.
+| Property | Value |
+|---|---|
+| Default exit | 22 layers, 768 dimensions |
+| Available depths | 3, 6, 11, 22 layers |
+| Available dimensions | 768, 512, 256, 128, 64 |
+| Pair capacity | 32,768 tokens total, including special tokens |
 
-## What it brings
+## Quick start
 
-- **More relevant results.** Compare each document with the meaning of the query.
-- **A choice of compute sizes.** Select from 20 options to balance quality and latency.
-- **Longer documents.** Score query–document pairs with up to 32K tokens combined.
+The model uses the public [Matryoshka2DReranker loader](https://github.com/vllm-project/semantic-router/blob/8f98c8adfa7940c4769854649f8f0959a14973a5/src/training/model_embeddings/mmbert_32k/reranker_model.py). Its 20 separately stored heads require this reader; generic `CrossEncoder` does not load this architecture.
 
-## At a glance
+With PyTorch, Transformers 4.57.6, safetensors and huggingface_hub installed, make the tested public source importable:
 
-| Evaluation | Previous model | Vela |
-|---|---:|---:|
-| Judged-pool ranking · 96 multilingual queries · nDCG@10 | 0.778 | **0.799** |
-| SciFact · 300 queries · nDCG@10 | 0.651 | **0.659** |
+```bash
+git clone --filter=blob:none https://github.com/vllm-project/semantic-router.git
+cd semantic-router
+git checkout 8f98c8adfa7940c4769854649f8f0959a14973a5
+export PYTHONPATH="$PWD"
+```
 
-Results use native FP16, the full model and fixed candidate documents. Gains vary by language; SciFact is a previously used benchmark. Long-document quality remains uneven: accuracy on the 32K end-position test decreased from 51.0% to 49.0%.
+```python
+import torch
+from huggingface_hub import snapshot_download
+from transformers import AutoTokenizer
+from src.training.model_embeddings.mmbert_32k.reranker_model import Matryoshka2DReranker
+
+path = snapshot_download("llm-semantic-router/Vela-1.0-Encoder-307M-Reranker")
+tokenizer = AutoTokenizer.from_pretrained(path)
+model = Matryoshka2DReranker.from_pretrained(
+    path, use_flash_attn=False, torch_dtype=torch.float32,
+).eval()
+inputs = tokenizer(
+    ["When does the library open?"],
+    ["The library opens in the morning."],
+    padding=True, truncation=False, return_tensors="pt",
+)
+if inputs["input_ids"].shape[1] > 32768:
+    raise ValueError("Query and passage exceed the combined token budget")
+with torch.inference_mode():
+    scores = model(**inputs, layer_idx=22, dim_idx=768)["logits"]
+print(scores)
+```
 
-## Quick start
+Select another supported `layer_idx` / `dim_idx` to use its trained head. A sigmoid preserves score ordering but is not established as a calibrated relevance probability. Merely returning an intermediate exit does not prove lower latency: the Python reader above runs the full encoder.
 
-Install PyTorch 2.10 for your device and the [inference dependencies](reproduction/requirements.txt), then run:
+## Measured performance
 
-```bash
-hf download llm-semantic-router/Vela-1.0-Encoder-307M-Reranker \
-  --local-dir vela-reranker --exclude 'onnx/*'
-python vela-reranker/reproduction/inference.py \
-  --task reranker --model vela-reranker \
-  --query 'How can I route requests to a cheaper model?' \
-  --text 'Semantic routing selects a suitable language model for each request.' \
-         'Gardening advice for growing tomatoes.'
-```
+FP32 comparison on the same known development candidate pools: 320 MIRACL queries across Arabic, Spanish, Japanese and Chinese; 128 NQ queries; 48 SciFact queries; 64 QASPER queries; and 288 controlled long-context pairs. All scores below use the full 22-layer/768-dimensional exit.
 
-Higher scores indicate greater relevance to the same query. Scores are not calibrated probabilities. The example runs on CPU.
+| Development measure | Previous Vela | This release |
+|---|---:|---:|
+| MIRACL nDCG@10 | 0.848966 | 0.819203 |
+| NQ nDCG@10 | 0.983537 | 0.979743 |
+| SciFact nDCG@10 | 0.889785 | 0.875239 |
+| QASPER nDCG@10 | 0.599072 | 0.416218 |
+| Controlled long-pair accuracy | 0.618056 | 0.600694 |
+| Controlled 4K pair accuracy | 0.611111 | 0.680556 |
+| 32K end-position pair accuracy · 24 pairs | 0.416667 | 0.500000 |
 
-## The Vela family
+The 4K and 32K end-position controls improve, while aggregate retrieval and several other context measures regress. This checkpoint was selected using development scores. These development scores are not independent test results, full-corpus search results or proof of broad natural long-document understanding.
 
-Choose the signals your router needs. Every model has a focused role.
+## Vela model family
 
-| Role | Models |
+| Model | Role |
 |---|---|
-| Understand requests | [Domain](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Domain) · [Feedback](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Feedback) · [Modality](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Modality) |
-| Detect risks | Guard¹ · [PII](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-PII) · Safety¹ · Hazard¹ |
-| Decide when to verify | [FactCheck](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-FactCheck) |
-| Retrieve context | [Embedding](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Embedding) · [Reranker](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Reranker) |
-| Build new capabilities | [Encoder](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M) |
-
-¹ Coming soon. Explore available models in the [Vela collection](https://huggingface.co/collections/llm-semantic-router/vela-10-router-models-6aa555ba70cc6997d6d67798).
+| [Base](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M) | Shared encoder foundation |
+| [Embedding](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Embedding) | Text representations and retrieval |
+| [Reranker](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Reranker) | Query–document relevance scoring |
+| [Domain](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Domain) | Request topic classification |
+| [Modality](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Modality) | Required output modality |
+| [Feedback](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Feedback) | User feedback classification |
+| [FactCheck](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-FactCheck) | External factual knowledge needs |
+| [PII](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-PII) | Sensitive entity spans |
+| [Guard](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Guard) | Prompt attack detection |
+| [Safety](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Safety) | General content risk |
+| [Hazard](https://huggingface.co/llm-semantic-router/Vela-1.0-Encoder-307M-Hazard) | Independent content risk categories |

THIRD_PARTY_NOTICES.md

@@ -1,14 +0,0 @@
-# Source and license notices
-
-Vela model weights and the Semantic Router code use Apache-2.0; see LICENSE. This does not change the terms or provenance of the datasets used to train or evaluate the model. Training and evaluation data must be obtained from their original providers. This package provides scripts, identifiers, hashes, and numerical results, and does not bundle source corpus text.
-
-| Source | Pinned input and use | Attribution and scope |
-|---|---|---|
-| Original mmBERT Reranker | `llm-semantic-router/mmbert-rerank-32k-2d-matryoshka@8022697d37c2602f06eff94cd6d83bb9f6ab1001`; original encoder and all ranking heads | [Original model](https://huggingface.co/llm-semantic-router/mmbert-rerank-32k-2d-matryoshka); Apache-2.0. Inherited pretraining exposures are not fully known. |
-| MIRACL | `miracl/miracl@ad2acf33f265b8fba92e5096c9d2cf569a82b067`; converted parquet snapshot, AR/ES/JA/ZH train and official dev | [MIRACL](https://github.com/project-miracl/miracl). Annotation/code licensing does not replace Wikipedia text attribution and applicable reuse terms. |
-| QASPER | `allenai/qasper@06806e4608976fc2fac0a090ac425d5b2b29caf4`; official train for natural-paper supervision, validation partitions for diagnostics | [QASPER](https://github.com/allenai/qasper); CC-BY-4.0. Retain paper/source attribution. Alternative papers are unjudged retrieval candidates. |
-| PAWS-X | `google-research-datasets/paws-x@4cd8187c404bda33cb1f62b49b001115862acf37`; historical training and six-language paraphrase diagnostics | [PAWS-X](https://github.com/google-research-datasets/paws/tree/master/pawsx). PAWS-Wiki-derived data; preserve the provider's dataset notice. The reusable loader removes empty and untranslated NS placeholders by default. |
-| STS-B | `sentence-transformers/stsb@ab7a5ac0e35aa22088bdcf23e7fd99b220e53308`; shared embedding retention/development pipeline | [STS benchmark](https://ixa2.si.ehu.eus/stswiki/index.php/STSbenchmark). Annotation and source-text terms differ; the complete corpus is not relabeled Apache-2.0. |
-| SciFact | `BeIR/scifact@cf10ab6856b15b0e670ef8ae5dae4e266c12d035`; evaluation only | David Wadden and colleagues, *Fact or Fiction: Verifying Scientific Claims* (EMNLP 2020). The [official component licenses](https://github.com/allenai/scifact/blob/68b98a56d93e0f9da0d2aab4e6c3294699a0f72e/LICENSE.md) are CC-BY-4.0 for claims/evidence annotations, ODC-By-1.0 for S2ORC abstracts, and Apache-2.0 for repository code. Attribution includes original article authors and S2ORC contributors. No SciFact corpus is redistributed here. |
-
-The source downloading script preserves README/license files and records the requested immutable revision plus content hashes. Reuse of a dataset is subject to that dataset's terms; evaluation files and derived numerical scores do not transfer ownership of source articles.

classification_heads.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:b6cb1a057445f5bd6da54823d5853a0d0c1568ea39bb076dce5fa04dfea93495
-size 7560529

onnx/fp16/layer-11-dim-128-ck.onnx → classification_heads.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23cb67b21462b49bb3052bc30caf257da586d29e46db89845628b070b45bced4
-size 77217
+oid sha256:040b619a7829035f87b34539c7d213b63e8a928b4f212cffe9cbcab426e84bec
+size 7537952

config.json

@@ -45,6 +45,14 @@
     "pooling": "cls",
     "version": 1
   },
+  "rope_scaling": {
+    "beta_fast": 32.0,
+    "beta_slow": 1.0,
+    "factor": 4.0,
+    "original_max_position_embeddings": 8192,
+    "rope_type": "yarn",
+    "truncate": true
+  },
   "sep_token_id": 1,
   "sparse_pred_ignore_index": -100,
   "sparse_prediction": false,

frozen-native-sha256.json

@@ -1,10 +0,0 @@
-{
-  "classification_heads.pt": "b6cb1a057445f5bd6da54823d5853a0d0c1568ea39bb076dce5fa04dfea93495",
-  "config.json": "5ffe270cb6d458fbdf6f223f86ae63bf21e72293b7f683dc1d0790dca1b0f864",
-  "matryoshka_config.json": "067d2e6cc6705d17e377b3a46d6aea48a932a9e246a2e554af074a33d245a0e3",
-  "model.safetensors": "86b7b4057f00246ce09d5efb6f78beebb132b50508aebe0055ac371726713dfc",
-  "special_tokens_map.json": "6e3204c5a4004719c185007a745d1940471a6fb02521cfa0a00306dbb6bc5903",
-  "tokenizer.json": "22fd4a60565f25fee8ebd2866aab07456dd46cbbeb45eaf506529ed8451ee58f",
-  "tokenizer_config.json": "fb401ced9a11236f919f50ff0850a0299b3b791d0caa48c8c6b2686ef51974a8",
-  "training_provenance.json": "fce2a66134a0504b56f0baab693cae9e0ecb13ee50520554659b3e5f578e534e"
-}

matryoshka_config.json

@@ -15,7 +15,11 @@
   "hidden_size": 768,
   "num_layers": 22,
   "pooling_strategy": "cls",
-  "has_final_norm": true,
-  "model_name_or_path": "llm-semantic-router/Vela-1.0-Encoder-307M-Reranker",
-  "model_revision": null
+  "representation_contract": {
+    "version": 1,
+    "intermediate_normalization": "final_norm",
+    "final_normalization": "final_norm",
+    "pooling": "cls",
+    "head_dtype": "float32"
+  }
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:86b7b4057f00246ce09d5efb6f78beebb132b50508aebe0055ac371726713dfc
+oid sha256:3cc023ed4c935d0a8201361c02dc3a543860d3798356c235fe9f9afb46b2ca29
 size 1227771776

onnx/fp16/layer-11-dim-128.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3770dd31cf882d6ef5f3f55924413602f1d2fd5dc3caac0ef3897983c6a0463d
-size 86043

onnx/fp16/layer-11-dim-256-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:7450d03a2a06df557e0940c5c591235086c4d7aea239a1cedc8eee804ab9e07c
-size 76838

onnx/fp16/layer-11-dim-256.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0e4df2e6b3f09ceb5e0427130b6ffcf173fb1e83193316480e19e3101fd9d490
-size 85664

onnx/fp16/layer-11-dim-512-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:382b2bdaffca33339bc8762757b085bda117949b41036256ca4349b69c84c51a
-size 76840

onnx/fp16/layer-11-dim-512.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c0638a644de9ad8840cac6dc8e35589e18724a963db038dc1f9a3552cbb4d08b
-size 85666

onnx/fp16/layer-11-dim-64-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:579bd57a62f04e8b4521f1a367e6748224f327ab1d9c1dd891cd8874a119217e
-size 76908

onnx/fp16/layer-11-dim-64.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f7cb1259d8d16457576039ef99f15781f39cbbdb6827fe9ed84ed80b6418ecc6
-size 85734

onnx/fp16/layer-11-dim-768-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:dfc0f61c3767246d55478b847dede149e084c9d683fd23c9795dda3feb9e5d82
-size 76454

onnx/fp16/layer-11-dim-768.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:22810eeb3144fe2dd39dae27954655e4f9fa6fabf20131c96997f917ee72e994
-size 85277

onnx/fp16/layer-22-dim-128-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:800e06cf290f99dea2614f1379a8fcbbd97810c9f86746ce57e3af714aab3ded
-size 148045

onnx/fp16/layer-22-dim-128.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3d227c44be829f7eaf2c2fe054203a33c0e691d0eea349892078c0b7c0c2ddf6
-size 166130

onnx/fp16/layer-22-dim-256-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:af081af8e15595fd7c374add040ce50f2fdabd6e6aa7f926fca40ff1d571ece4
-size 147666

onnx/fp16/layer-22-dim-256.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:096abccda15e50d04199ba4bfea75d0c5efb0adbdb19e3115041b6dbea2aad6b
-size 165751

onnx/fp16/layer-22-dim-512-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:3cc366b7d1ca50926490b663a943287a19bbe71e37bb56daa2b7e2b3ec215768
-size 147668

onnx/fp16/layer-22-dim-512.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2517310224da947e8187f76ef53aaa31a1a55915a332c025637a48bd5de8b53f
-size 165753

onnx/fp16/layer-22-dim-64-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:56bc9c9dae7c4db6d2718b6cf9247193d96d1c3faa6c2a8f8d34becc3755ad3a
-size 147736

onnx/fp16/layer-22-dim-64.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:02e18a742086874c058c19670a85ca98972c43e8ac08020b32b73920389f8ef5
-size 165821

onnx/fp16/layer-22-dim-768-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e1d96dac7d4f80fb2e83475c5988c3996f47b150b3758a355b27b01b8163add2
-size 147084

onnx/fp16/layer-22-dim-768.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:673e336f60b631bd580d50ef4db86fd6b8c266508c74f972989c5d7bf6e881ee
-size 165166

onnx/fp16/layer-3-dim-128-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d84968d49b2f171e262a51285068168c6a51a4c90c32fe86637e7a8d20721f68
-size 26243

onnx/fp16/layer-3-dim-128.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ff320ef2fa7ae956dee40320eee96f9e3c1cafa97f192e74ed686766bd0759f4
-size 28577

onnx/fp16/layer-3-dim-256-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:11b637e3dd042821d74aa6e2ae0fb817af3bffc81f04856127f3c16356fbd0c8
-size 25864

onnx/fp16/layer-3-dim-256.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:30ea7650de17e31177afd13efd7e555cf1ec17f070c9b8af79b71405953db16f
-size 28198

onnx/fp16/layer-3-dim-512-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9881d9d870bf83801a9675b336c18e93ca6d0ca31f738f4ccc916219d74da820
-size 25866

onnx/fp16/layer-3-dim-512.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:85de5ad9a9b2b86f59bc5424f8f77bc4c5583ca63bad018e25a38201cf672bb0
-size 28200

onnx/fp16/layer-3-dim-64-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0e783eef5113020d0b9b0b776b3fdca775d04fad948ac59cb194c3215801ddc2
-size 25933

onnx/fp16/layer-3-dim-64.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:eb0ac3efa0309872a4fef61557c7c3be24159c6f3def44b37d8e463d47b2a76d
-size 28267

onnx/fp16/layer-3-dim-768-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:536c0f7a0036d0d95b663a1095b53c14add9a403f23f7d6a1dd933ace2dbb2e6
-size 25652

onnx/fp16/layer-3-dim-768.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:48743a084e494a39f35eacbe95cfef537c9227685d8b38cafcf04cbacc60971b
-size 27984

onnx/fp16/layer-6-dim-128-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2fd3f6991370c6ef8f47bd791249f6059083f3d8e670b9a604a865c4ed5f5d27
-size 45169

onnx/fp16/layer-6-dim-128.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4aae895ed61dca617a1a43840f7bb3dcf1d43e333eca4c507c9f9d3d43d0010d
-size 49891

onnx/fp16/layer-6-dim-256-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:5537d051bf84bfa68e0dc01a622a6a584fc28ad96d030e07377f5175c8e922be
-size 44790

onnx/fp16/layer-6-dim-256.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:2e1790d8fce657b341b7be1bb16af93b402c8bdf8cec478ef87a297cb4248ea9
-size 49512

onnx/fp16/layer-6-dim-512-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:f8c062eb1d2378c9743e8a7b8b206384abbd3f312c012862f873e0e68b0b1193
-size 44792

onnx/fp16/layer-6-dim-512.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:25d61c845753bce34b5f8c493d68fd41187f6f83ff95fff573c87f0abdae26f9
-size 49514

onnx/fp16/layer-6-dim-64-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:fe6f82173b9961ab65a5e71c1cdd5ac1634238309f31987c1b5d9c6346a12f93
-size 44859

onnx/fp16/layer-6-dim-64.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9a62fbd19bfa1d62a4711ea6eda382c22100d6bb25296ce93c11a445f26f005b
-size 49581

onnx/fp16/layer-6-dim-768-ck.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:23b06b498b0fac0b81658c12fd277c5e29e60c7a9966c7550cd4c71c26ae462e
-size 44542

onnx/fp16/layer-6-dim-768.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:4a1cb2a1a92a350405fea6cc1da337d8515a862e103dca50d2f3c84b65db29cd
-size 49262

onnx/fp16/weights.data

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ccff229c19e05d1cf05edf2a924ea20c556288b339c2d44d21038dceb73e987e
-size 621407744

onnx/fp32/layer-11-dim-128.onnx

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:567a4eb8105a60a14f99b8b3aa638c30985b83bdc35350b22d1aa74bc2527a07
-size 85595