Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+variants/cpu-int4/inference/frozen_base.onnx.data filter=lfs diff=lfs merge=lfs -text
+variants/cpu-int4/train/checkpoint filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,49 @@
+---
+base_model: sentence-transformers/all-MiniLM-L6-v2
+library_name: mobiletransformers
+pipeline_tag: text-classification
+tags:
+  - mobiletransformers
+  - onnx
+  - on-device
+  - android
+  - lora
+  - int4
+---
+
+# sentence-transformers/all-MiniLM-L6-v2 — MobileTransformers package
+
+On-device (Android) package exported from **sentence-transformers/all-MiniLM-L6-v2** with MobileTransformers.
+
+## What this package can do
+- `core` — shared files every other group needs
+- `inference` — generate or score on device
+- `train` — **fine-tune on device**, then merge the adapter back into the base weights
+- `rag` — retrieve over documents you ingest, and ground answers in them
+
+## Fine-tuning method
+- **lora** — LoRA — low-rank adapters on the attention projections.
+- Rank: `8`
+- Adapted modules: `query`, `value`
+
+## Provenance
+- Base model: `sentence-transformers/all-MiniLM-L6-v2`
+- Selected task: `text-classification`
+- Quantization: int4
+- Toolchain: optimum-onnx 0.1.0, transformers 4.57.6, ort-training 1.23.0+cpu
+
+## Licenses
+- Framework: not declared in this package — see the repository
+- Base model weights: see the base model above (this package redistributes an export of those weights, so their terms govern its contents)
+
+## Android runtime
+- Minimum API: 28
+- Required ABIs: any
+
+## Variants
+
+| id | EP | quant | engines | features | min API | rec. RAM (MB) |
+| --- | --- | --- | --- | --- | --- | --- |
+| cpu-int4 | cpu | int4 | native | core, inference, train, rag | 28 | — |
+
+Default variant: `cpu-int4`.

mobiletransformers_manifest.json

@@ -0,0 +1,205 @@
+{
+  "androidRuntime": {
+    "minimumAndroidApi": 28,
+    "recommendedDeviceMemoryMb": null,
+    "requiredAbis": []
+  },
+  "architectures": [
+    "bert"
+  ],
+  "artifactFormatVersion": 1,
+  "baseModelId": "sentence-transformers/all-MiniLM-L6-v2",
+  "defaultVariant": "cpu-int4",
+  "downloadPlan": {
+    "cpu-int4": {
+      "checksums": [
+        "variants/cpu-int4/checksums.json"
+      ],
+      "core": [
+        "mobiletransformers_manifest.json",
+        "shared/tokenizer/**",
+        "shared/chat_template.jinja",
+        "shared/config.json",
+        "shared/generation_config.json"
+      ],
+      "genai": [],
+      "inference": [
+        "variants/cpu-int4/inference/**"
+      ],
+      "rag": [
+        "variants/cpu-int4/embedding/**"
+      ],
+      "train": [
+        "variants/cpu-int4/train/**"
+      ]
+    }
+  },
+  "exportedAt": "",
+  "fileSizes": {
+    "optimum/export_report.json": 758,
+    "optimum/supported_tasks.json": 28,
+    "shared/tokenizer/mobiletransformers_tokenizer_config.json": 296,
+    "shared/tokenizer/special_tokens_map.json": 695,
+    "shared/tokenizer/tokenizer.json": 711661,
+    "shared/tokenizer/tokenizer_config.json": 1464,
+    "shared/tokenizer/vocab.txt": 231508,
+    "variants/cpu-int4/checksums.json": 6585,
+    "variants/cpu-int4/embedding/embedding_model.onnx": 90392834,
+    "variants/cpu-int4/embedding/rag_config.json": 346,
+    "variants/cpu-int4/embedding/tokenizer/special_tokens_map.json": 695,
+    "variants/cpu-int4/embedding/tokenizer/tokenizer.json": 711661,
+    "variants/cpu-int4/embedding/tokenizer/tokenizer_config.json": 1464,
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin": 589824,
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin.sha256": 65,
+    "variants/cpu-int4/inference/config.json": 611,
+    "variants/cpu-int4/inference/frozen_base.onnx.data": 83778056,
+    "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": 811,
+    "variants/cpu-int4/inference/model.onnx": 110896,
+    "variants/cpu-int4/inference/optimum_config.json": 378,
+    "variants/cpu-int4/inference/special_tokens_map.json": 695,
+    "variants/cpu-int4/inference/tokenizer.json": 711661,
+    "variants/cpu-int4/inference/tokenizer_config.json": 1464,
+    "variants/cpu-int4/inference/vocab.txt": 231508,
+    "variants/cpu-int4/inference/weight_handoff_map.json": 18707,
+    "variants/cpu-int4/train/checkpoint": 28875552,
+    "variants/cpu-int4/train/eval_model.onnx": 250699,
+    "variants/cpu-int4/train/optimizer_model.onnx": 538,
+    "variants/cpu-int4/train/trainable_parameters.json": 2762,
+    "variants/cpu-int4/train/training_config.json": 7539,
+    "variants/cpu-int4/train/training_model.onnx": 373568
+  },
+  "license": {
+    "baseModelWeights": null,
+    "framework": null,
+    "noticeFile": null
+  },
+  "minReaderVersion": "1.0",
+  "mobiletransformersVersion": "0.1.0",
+  "onnxRuntimeGenAIVersion": null,
+  "onnxRuntimeTrainingVersion": "1.23.0+cpu",
+  "optimumOnnxVersion": "0.1.0",
+  "peftMethods": [
+    "lora"
+  ],
+  "quantization": [
+    "int4"
+  ],
+  "requiredFiles": [
+    "mobiletransformers_manifest.json",
+    "shared/tokenizer/tokenizer.json",
+    "variants/cpu-int4/inference/model.onnx"
+  ],
+  "schemaVersion": "1.0",
+  "selectedTask": "text-classification",
+  "sha256": {
+    "optimum/export_report.json": "789837f0dc69497c044f7804266ce70329b5acdd0b9260a299c2e03b2150e195",
+    "optimum/supported_tasks.json": "00b8729cc70217e04b8951a7b059687aee182c90bb0e9254efec11d74988cd42",
+    "shared/tokenizer/mobiletransformers_tokenizer_config.json": "62dc6dc13bbe7e36bb93fa0d7859e3f90201410304eaf8a8ca91335b9ce4b918",
+    "shared/tokenizer/special_tokens_map.json": "5d5b662e421ea9fac075174bb0688ee0d9431699900b90662acd44b2a350503a",
+    "shared/tokenizer/tokenizer.json": "da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0",
+    "shared/tokenizer/tokenizer_config.json": "ccb4eb21a03e1442ee5c3f85431b9c307960a04579942537d74778dc8080a48c",
+    "shared/tokenizer/vocab.txt": "07eced375cec144d27c900241f3e339478dec958f92fddbc551f295c992038a3",
+    "variants/cpu-int4/checksums.json": "6e0726657fe1193dae16b320f48cbe1b5cd58ab3d3bf4bd98351db88d35883f1",
+    "variants/cpu-int4/embedding/embedding_model.onnx": "7c46645e1dcec3d3a27d80f21fcad81ab307317bf3d0580eb8ea6779fa83cc96",
+    "variants/cpu-int4/embedding/rag_config.json": "99088538b87ca7e78613f328c4c7eb801cf4dc9b2443ddb6d3b9b59acf486f92",
+    "variants/cpu-int4/embedding/tokenizer/special_tokens_map.json": "5d5b662e421ea9fac075174bb0688ee0d9431699900b90662acd44b2a350503a",
+    "variants/cpu-int4/embedding/tokenizer/tokenizer.json": "da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0",
+    "variants/cpu-int4/embedding/tokenizer/tokenizer_config.json": "ccb4eb21a03e1442ee5c3f85431b9c307960a04579942537d74778dc8080a48c",
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin": "e009749b75204bb57ade70357050c824f6a93d9b1f920405cc77ce988ed74e4f",
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin.sha256": "252a25316a60670ee5c788338517203772198f016ec46b3ed6573290655be36c",
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin": "1a99fa38c68e9f893e79d1439a25f90902158f0379ca4ab1e66f27cd3e4fbd0f",
+    "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin.sha256": "53765830cb7c6e8ddc656ada7f231ec3c5a11e9bee2d5e0566efed11436abc55",
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin": "3e8b8e464be41cf14af6371fea3d7fc3abf0b38b6031994aa44b4e5fffb34fe4",
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin.sha256": "fa1e53ad2e02c039edc44c944a1fa6c470d2aef0002533f0206c2a0e034dbdd7",
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin": "54470c78602372ab240b6f88815f7c8e86ece20ddd7066ca7c8e342dfa7d51bd",
+    "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin.sha256": "b209496517feff85bfd286e7287c0c9381d501e36585a0362b4ba88edb629e7b",
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin": "b21671817e578d43e5fea24c2bd15686bb951e65dcf69054cf6d76747271a40b",
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin.sha256": "c0d24d803bb92f6edcecdd305a50dc7bf7a8d5c2535e8f0e50bcff629a612cbe",
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin": "4dd4a89014a92c780cbe209eb778a052e27c5c303a246a9f9b3233ec3a6c2390",
+    "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin.sha256": "40ef2b5cc5dcb0d28efd7f2e0500ae9e94e2acbfc98cf1e31f2c4978ae25b5a1",
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin": "bbb7a4be247912940098e59a250eb63a7eaa24ff1abe2ddcd8f463372b9ddd24",
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin.sha256": "36ffffc6dbf9990d9a3eb05645cbaf05419cb0710429c072c2826514b5a24577",
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin": "97f887ce9906ac2eb845e7a6ae11ac3bdb73de78bffc182337324f6a92c94da2",
+    "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin.sha256": "fe896676a74cad17d28e64df42db042c0082966597118754723ce5ef8b95fae0",
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin": "8ea4af517bd57ff2ad5841ab0c8256fc8ed7981ecbb739f504f642b54ad9e727",
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin.sha256": "fbdda2a28ddfd5c2ffd0a8e55b2ac64b56862acc99226c35c839692d5173e6d6",
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin": "34111787d5ee0330d9ea6d3e8577f32746aab2ae4c61f2ddcf8c287a8e17269a",
+    "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin.sha256": "7f1d26e09b97106e3d825565aa3a1ccb233479f6dfb8cc31c2805e878cb89843",
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin": "97d659bc916216396647f25e5d60039190eafe0d3ad3908cba326299c1dcf0c0",
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin.sha256": "d83d46dbc5ad9a27919a8c6b91218270e043888e887ecbb988aeb77412e8fbc0",
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin": "71fcbee801f47cad8b13e6f19286d988ac44e778d8616867d8b1b280f6b7e35a",
+    "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin.sha256": "a688cc07a80e888c1260a5d9087d37fb1cdb717d0c20a23aabb4f36617952d6a",
+    "variants/cpu-int4/inference/config.json": "4599e8a5d74ed192b70919aa02124c2d68580070b22e89db956c0353618bccd9",
+    "variants/cpu-int4/inference/frozen_base.onnx.data": "bd46ab7db78e5677a6e581370c3691791cb556623acbd3f7d011938da28bf331",
+    "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": "953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c",
+    "variants/cpu-int4/inference/model.onnx": "cf205aaf284fac7b7e8c2a8e02634a926b1e7af434da7d01d287450bb35ed032",
+    "variants/cpu-int4/inference/optimum_config.json": "3dae0721ec58df335f0b6acd0bda466a5637b1686bb052c43ea8c2b3d9a757da",
+    "variants/cpu-int4/inference/special_tokens_map.json": "5d5b662e421ea9fac075174bb0688ee0d9431699900b90662acd44b2a350503a",
+    "variants/cpu-int4/inference/tokenizer.json": "da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0",
+    "variants/cpu-int4/inference/tokenizer_config.json": "ccb4eb21a03e1442ee5c3f85431b9c307960a04579942537d74778dc8080a48c",
+    "variants/cpu-int4/inference/vocab.txt": "07eced375cec144d27c900241f3e339478dec958f92fddbc551f295c992038a3",
+    "variants/cpu-int4/inference/weight_handoff_map.json": "8f158641089032f0b9beaf8dfe86d0601d67b648528fd2474e373d83f7a0a5c3",
+    "variants/cpu-int4/train/checkpoint": "9fb3cfb3bc0b8f0ae8f0019531e138858069eba2379bf81dc17ff80552825225",
+    "variants/cpu-int4/train/eval_model.onnx": "2e98cd22e60dbb87b4e3aa26bf61078bcf319d358a9951c1e15a7ee4618cb3bc",
+    "variants/cpu-int4/train/optimizer_model.onnx": "8ed3b548bce52bf25c2154b1452314b3c48d079e28c52822dd348034f59dc054",
+    "variants/cpu-int4/train/trainable_parameters.json": "6d39ef39887770fe5e2b1f121b3077df0911142f186e57ac849808806f772654",
+    "variants/cpu-int4/train/training_config.json": "790901caaa951e9a8f89cc63f109b680007c1cc707920a87576b52a01c7130ca",
+    "variants/cpu-int4/train/training_model.onnx": "56bf04b82b8c796a2a36e8a79e09454776424857f746036bfdd9c58c9ed70875"
+  },
+  "supportedTasks": [
+    "text-classification"
+  ],
+  "trainableParameterCount": 73728,
+  "trainingParameterCount": 22861497,
+  "transformersVersion": "4.57.6",
+  "trustRemoteCode": false,
+  "variants": [
+    {
+      "abi": null,
+      "executionProvider": "cpu",
+      "features": [
+        "core",
+        "inference",
+        "train",
+        "rag"
+      ],
+      "id": "cpu-int4",
+      "minimumAndroidApi": 28,
+      "paths": {
+        "embedding": "variants/cpu-int4/embedding",
+        "inference": "variants/cpu-int4/inference",
+        "tokenizer": "shared/tokenizer",
+        "train": "variants/cpu-int4/train"
+      },
+      "quantization": "int4",
+      "recommendedDeviceMemoryMb": null,
+      "supportedEngines": [
+        "native"
+      ],
+      "weightHandoff": "variants/cpu-int4/inference/weight_handoff_map.json"
+    }
+  ],
+  "weightHandoff": "variants/cpu-int4/inference/weight_handoff_map.json"
+}

optimum/export_report.json

@@ -0,0 +1,34 @@
+{
+  "androidRuntime": {
+    "minimumAndroidApi": 28,
+    "recommendedDeviceMemoryMb": null,
+    "requiredAbis": []
+  },
+  "architectures": [
+    "bert"
+  ],
+  "license": {
+    "baseModelWeights": null,
+    "framework": null,
+    "noticeFile": null
+  },
+  "mobiletransformersVersion": "0.1.0",
+  "onnxRuntimeTrainingVersion": "1.23.0+cpu",
+  "optimumOnnxVersion": "0.1.0",
+  "peftMethods": [
+    "lora"
+  ],
+  "quantization": [
+    "int4"
+  ],
+  "selectedTask": "text-classification",
+  "sourceParameterCount": 22713986,
+  "supportedTasks": [
+    "text-classification"
+  ],
+  "trainableParameterCount": 73728,
+  "trainingParameterCount": 22861497,
+  "trainingQuantizedParameterCount": 20765211,
+  "transformersVersion": "4.57.6",
+  "trustRemoteCode": false
+}

optimum/supported_tasks.json

@@ -0,0 +1,3 @@
+[
+  "text-classification"
+]

shared/tokenizer/mobiletransformers_tokenizer_config.json

@@ -0,0 +1,13 @@
+{
+    "model": {
+        "bos_token_id": null,
+        "context_length": 512,
+        "num_attention_heads": 12,
+        "num_hidden_layers": 6,
+        "num_key_value_heads": 12,
+        "eos_token_id": null,
+        "pad_token_id": 0,
+        "type": "bert",
+        "vocab_size": 30522
+    }
+}

shared/tokenizer/special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

shared/tokenizer/tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "max_length": 128,
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_to_multiple_of": null,
+  "pad_token": "[PAD]",
+  "pad_token_type_id": 0,
+  "padding_side": "right",
+  "sep_token": "[SEP]",
+  "stride": 0,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "truncation_side": "right",
+  "truncation_strategy": "longest_first",
+  "unk_token": "[UNK]"
+}

variants/cpu-int4/checksums.json

@@ -0,0 +1,48 @@
+{
+  "variants/cpu-int4/checksums.json": "e4a42e8a583cd6a5ac8066f6d35b08133dd3c5099123c9e745f9165146122acf",
+  "variants/cpu-int4/embedding/embedding_model.onnx": "7c46645e1dcec3d3a27d80f21fcad81ab307317bf3d0580eb8ea6779fa83cc96",
+  "variants/cpu-int4/embedding/rag_config.json": "99088538b87ca7e78613f328c4c7eb801cf4dc9b2443ddb6d3b9b59acf486f92",
+  "variants/cpu-int4/embedding/tokenizer/special_tokens_map.json": "5d5b662e421ea9fac075174bb0688ee0d9431699900b90662acd44b2a350503a",
+  "variants/cpu-int4/embedding/tokenizer/tokenizer.json": "da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0",
+  "variants/cpu-int4/embedding/tokenizer/tokenizer_config.json": "ccb4eb21a03e1442ee5c3f85431b9c307960a04579942537d74778dc8080a48c",
+  "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin": "e009749b75204bb57ade70357050c824f6a93d9b1f920405cc77ce988ed74e4f",
+  "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin.sha256": "252a25316a60670ee5c788338517203772198f016ec46b3ed6573290655be36c",
+  "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin": "1a99fa38c68e9f893e79d1439a25f90902158f0379ca4ab1e66f27cd3e4fbd0f",
+  "variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin.sha256": "53765830cb7c6e8ddc656ada7f231ec3c5a11e9bee2d5e0566efed11436abc55",
+  "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin": "3e8b8e464be41cf14af6371fea3d7fc3abf0b38b6031994aa44b4e5fffb34fe4",
+  "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin.sha256": "fa1e53ad2e02c039edc44c944a1fa6c470d2aef0002533f0206c2a0e034dbdd7",
+  "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin": "54470c78602372ab240b6f88815f7c8e86ece20ddd7066ca7c8e342dfa7d51bd",
+  "variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin.sha256": "b209496517feff85bfd286e7287c0c9381d501e36585a0362b4ba88edb629e7b",
+  "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin": "b21671817e578d43e5fea24c2bd15686bb951e65dcf69054cf6d76747271a40b",
+  "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin.sha256": "c0d24d803bb92f6edcecdd305a50dc7bf7a8d5c2535e8f0e50bcff629a612cbe",
+  "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin": "4dd4a89014a92c780cbe209eb778a052e27c5c303a246a9f9b3233ec3a6c2390",
+  "variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin.sha256": "40ef2b5cc5dcb0d28efd7f2e0500ae9e94e2acbfc98cf1e31f2c4978ae25b5a1",
+  "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin": "bbb7a4be247912940098e59a250eb63a7eaa24ff1abe2ddcd8f463372b9ddd24",
+  "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin.sha256": "36ffffc6dbf9990d9a3eb05645cbaf05419cb0710429c072c2826514b5a24577",
+  "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin": "97f887ce9906ac2eb845e7a6ae11ac3bdb73de78bffc182337324f6a92c94da2",
+  "variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin.sha256": "fe896676a74cad17d28e64df42db042c0082966597118754723ce5ef8b95fae0",
+  "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin": "8ea4af517bd57ff2ad5841ab0c8256fc8ed7981ecbb739f504f642b54ad9e727",
+  "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin.sha256": "fbdda2a28ddfd5c2ffd0a8e55b2ac64b56862acc99226c35c839692d5173e6d6",
+  "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin": "34111787d5ee0330d9ea6d3e8577f32746aab2ae4c61f2ddcf8c287a8e17269a",
+  "variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin.sha256": "7f1d26e09b97106e3d825565aa3a1ccb233479f6dfb8cc31c2805e878cb89843",
+  "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin": "97d659bc916216396647f25e5d60039190eafe0d3ad3908cba326299c1dcf0c0",
+  "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin.sha256": "d83d46dbc5ad9a27919a8c6b91218270e043888e887ecbb988aeb77412e8fbc0",
+  "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin": "71fcbee801f47cad8b13e6f19286d988ac44e778d8616867d8b1b280f6b7e35a",
+  "variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin.sha256": "a688cc07a80e888c1260a5d9087d37fb1cdb717d0c20a23aabb4f36617952d6a",
+  "variants/cpu-int4/inference/config.json": "4599e8a5d74ed192b70919aa02124c2d68580070b22e89db956c0353618bccd9",
+  "variants/cpu-int4/inference/frozen_base.onnx.data": "bd46ab7db78e5677a6e581370c3691791cb556623acbd3f7d011938da28bf331",
+  "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": "953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c",
+  "variants/cpu-int4/inference/model.onnx": "cf205aaf284fac7b7e8c2a8e02634a926b1e7af434da7d01d287450bb35ed032",
+  "variants/cpu-int4/inference/optimum_config.json": "3dae0721ec58df335f0b6acd0bda466a5637b1686bb052c43ea8c2b3d9a757da",
+  "variants/cpu-int4/inference/special_tokens_map.json": "5d5b662e421ea9fac075174bb0688ee0d9431699900b90662acd44b2a350503a",
+  "variants/cpu-int4/inference/tokenizer.json": "da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0",
+  "variants/cpu-int4/inference/tokenizer_config.json": "ccb4eb21a03e1442ee5c3f85431b9c307960a04579942537d74778dc8080a48c",
+  "variants/cpu-int4/inference/vocab.txt": "07eced375cec144d27c900241f3e339478dec958f92fddbc551f295c992038a3",
+  "variants/cpu-int4/inference/weight_handoff_map.json": "8f158641089032f0b9beaf8dfe86d0601d67b648528fd2474e373d83f7a0a5c3",
+  "variants/cpu-int4/train/checkpoint": "9fb3cfb3bc0b8f0ae8f0019531e138858069eba2379bf81dc17ff80552825225",
+  "variants/cpu-int4/train/eval_model.onnx": "2e98cd22e60dbb87b4e3aa26bf61078bcf319d358a9951c1e15a7ee4618cb3bc",
+  "variants/cpu-int4/train/optimizer_model.onnx": "8ed3b548bce52bf25c2154b1452314b3c48d079e28c52822dd348034f59dc054",
+  "variants/cpu-int4/train/trainable_parameters.json": "6d39ef39887770fe5e2b1f121b3077df0911142f186e57ac849808806f772654",
+  "variants/cpu-int4/train/training_config.json": "790901caaa951e9a8f89cc63f109b680007c1cc707920a87576b52a01c7130ca",
+  "variants/cpu-int4/train/training_model.onnx": "56bf04b82b8c796a2a36e8a79e09454776424857f746036bfdd9c58c9ed70875"
+}

variants/cpu-int4/embedding/embedding_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c46645e1dcec3d3a27d80f21fcad81ab307317bf3d0580eb8ea6779fa83cc96
+size 90392834

variants/cpu-int4/embedding/rag_config.json

@@ -0,0 +1,13 @@
+{
+  "chunkOverlap": 50,
+  "chunkSize": 512,
+  "embeddingDimension": 384,
+  "embeddingModelId": "sentence-transformers/all-MiniLM-L6-v2",
+  "indexingMode": "precompute",
+  "maxTextLength": 1024,
+  "minScore": 0.0,
+  "onnxName": "embedding_model",
+  "repoName": "sentence-transformers__all-MiniLM-L6-v2",
+  "searchType": "semantic",
+  "topK": 10
+}

variants/cpu-int4/embedding/tokenizer/special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

variants/cpu-int4/embedding/tokenizer/tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "max_length": 128,
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_to_multiple_of": null,
+  "pad_token": "[PAD]",
+  "pad_token_type_id": 0,
+  "padding_side": "right",
+  "sep_token": "[SEP]",
+  "stride": 0,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "truncation_side": "right",
+  "truncation_strategy": "longest_first",
+  "unk_token": "[UNK]"
+}

variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e009749b75204bb57ade70357050c824f6a93d9b1f920405cc77ce988ed74e4f
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+e009749b75204bb57ade70357050c824f6a93d9b1f920405cc77ce988ed74e4f

variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a99fa38c68e9f893e79d1439a25f90902158f0379ca4ab1e66f27cd3e4fbd0f
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.0.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+1a99fa38c68e9f893e79d1439a25f90902158f0379ca4ab1e66f27cd3e4fbd0f

variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e8b8e464be41cf14af6371fea3d7fc3abf0b38b6031994aa44b4e5fffb34fe4
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+3e8b8e464be41cf14af6371fea3d7fc3abf0b38b6031994aa44b4e5fffb34fe4

variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:54470c78602372ab240b6f88815f7c8e86ece20ddd7066ca7c8e342dfa7d51bd
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.1.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+54470c78602372ab240b6f88815f7c8e86ece20ddd7066ca7c8e342dfa7d51bd

variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b21671817e578d43e5fea24c2bd15686bb951e65dcf69054cf6d76747271a40b
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+b21671817e578d43e5fea24c2bd15686bb951e65dcf69054cf6d76747271a40b

variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dd4a89014a92c780cbe209eb778a052e27c5c303a246a9f9b3233ec3a6c2390
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.2.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+4dd4a89014a92c780cbe209eb778a052e27c5c303a246a9f9b3233ec3a6c2390

variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbb7a4be247912940098e59a250eb63a7eaa24ff1abe2ddcd8f463372b9ddd24
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+bbb7a4be247912940098e59a250eb63a7eaa24ff1abe2ddcd8f463372b9ddd24

variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97f887ce9906ac2eb845e7a6ae11ac3bdb73de78bffc182337324f6a92c94da2
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.3.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+97f887ce9906ac2eb845e7a6ae11ac3bdb73de78bffc182337324f6a92c94da2

variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ea4af517bd57ff2ad5841ab0c8256fc8ed7981ecbb739f504f642b54ad9e727
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+8ea4af517bd57ff2ad5841ab0c8256fc8ed7981ecbb739f504f642b54ad9e727

variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34111787d5ee0330d9ea6d3e8577f32746aab2ae4c61f2ddcf8c287a8e17269a
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.4.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+34111787d5ee0330d9ea6d3e8577f32746aab2ae4c61f2ddcf8c287a8e17269a

variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97d659bc916216396647f25e5d60039190eafe0d3ad3908cba326299c1dcf0c0
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.query.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+97d659bc916216396647f25e5d60039190eafe0d3ad3908cba326299c1dcf0c0

variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71fcbee801f47cad8b13e6f19286d988ac44e778d8616867d8b1b280f6b7e35a
+size 589824

variants/cpu-int4/inference/bert.encoder.layer.5.attention.self.value.MatMul.weight.bin.sha256

@@ -0,0 +1 @@
+71fcbee801f47cad8b13e6f19286d988ac44e778d8616867d8b1b280f6b7e35a

variants/cpu-int4/inference/config.json

@@ -0,0 +1,25 @@
+{
+  "architectures": [
+    "BertModel"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "dtype": "float32",
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 384,
+  "initializer_range": 0.02,
+  "intermediate_size": 1536,
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 6,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "transformers_version": "4.57.6",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

variants/cpu-int4/inference/frozen_base.onnx.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd46ab7db78e5677a6e581370c3691791cb556623acbd3f7d011938da28bf331
+size 83778056

variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c
+size 811

variants/cpu-int4/inference/model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf205aaf284fac7b7e8c2a8e02634a926b1e7af434da7d01d287450bb35ed032
+size 110896

variants/cpu-int4/inference/optimum_config.json

@@ -0,0 +1,17 @@
+{
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1"
+  },
+  "inferenceGraphPrecision": "float32",
+  "modelId": "sentence-transformers/all-MiniLM-L6-v2",
+  "modelType": "bert",
+  "optimumOnnxVersion": "0.1.0",
+  "quantization": "int4",
+  "supportedEngines": [
+    "native"
+  ],
+  "task": "text-classification",
+  "transformersVersion": "4.57.6",
+  "trustRemoteCode": false
+}

variants/cpu-int4/inference/special_tokens_map.json

@@ -0,0 +1,37 @@
+{
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

variants/cpu-int4/inference/tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "max_length": 128,
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_to_multiple_of": null,
+  "pad_token": "[PAD]",
+  "pad_token_type_id": 0,
+  "padding_side": "right",
+  "sep_token": "[SEP]",
+  "stride": 0,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "truncation_side": "right",
+  "truncation_strategy": "longest_first",
+  "unk_token": "[UNK]"
+}

variants/cpu-int4/inference/weight_handoff_map.json

@@ -0,0 +1,652 @@
+{
+  "engines": [
+    "native",
+    "genai"
+  ],
+  "entries": [
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.0.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.0.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.0.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.0.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.0.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.0.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "e009749b75204bb57ade70357050c824f6a93d9b1f920405cc77ce988ed74e4f"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.0.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.0.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.0.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.0.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.0.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.0.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.0.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "1a99fa38c68e9f893e79d1439a25f90902158f0379ca4ab1e66f27cd3e4fbd0f"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.0.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.1.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.1.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.1.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.1.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.1.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.1.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "3e8b8e464be41cf14af6371fea3d7fc3abf0b38b6031994aa44b4e5fffb34fe4"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.1.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.1.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.1.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.1.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.1.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.1.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.1.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "54470c78602372ab240b6f88815f7c8e86ece20ddd7066ca7c8e342dfa7d51bd"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.1.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.2.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.2.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.2.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.2.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.2.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.2.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "b21671817e578d43e5fea24c2bd15686bb951e65dcf69054cf6d76747271a40b"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.2.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.2.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.2.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.2.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.2.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.2.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.2.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "4dd4a89014a92c780cbe209eb778a052e27c5c303a246a9f9b3233ec3a6c2390"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.2.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.3.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.3.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.3.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.3.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.3.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.3.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "bbb7a4be247912940098e59a250eb63a7eaa24ff1abe2ddcd8f463372b9ddd24"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.3.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.3.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.3.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.3.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.3.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.3.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.3.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "97f887ce9906ac2eb845e7a6ae11ac3bdb73de78bffc182337324f6a92c94da2"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.3.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.4.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.4.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.4.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.4.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.4.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.4.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "8ea4af517bd57ff2ad5841ab0c8256fc8ed7981ecbb739f504f642b54ad9e727"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.4.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.4.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.4.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.4.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.4.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.4.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.4.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "34111787d5ee0330d9ea6d3e8577f32746aab2ae4c61f2ddcf8c287a8e17269a"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.4.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.5.attention.self.query.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.5.attention.self.query.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.5.attention.self.query.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.5.attention.self.query.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.5.attention.self.query.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.5.attention.self.query.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "97d659bc916216396647f25e5d60039190eafe0d3ad3908cba326299c1dcf0c0"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.5.attention.self.query.base_layer",
+      "transposePolicy": "no_transpose"
+    },
+    {
+      "adapterDtypes": {
+        "adapter_A": "float32",
+        "adapter_B": "float32"
+      },
+      "adapterShapes": {
+        "adapter_A": [
+          8,
+          384
+        ],
+        "adapter_B": [
+          384,
+          8
+        ]
+      },
+      "checkpointNames": {
+        "adapter_A": "base_model.model.bert.encoder.layer.5.attention.self.value.lora_A.lora",
+        "adapter_B": "base_model.model.bert.encoder.layer.5.attention.self.value.lora_B.lora",
+        "weight": "base_model.model.bert.encoder.layer.5.attention.self.value.base_layer.weight"
+      },
+      "dtype": "float32",
+      "externalDataLocation": {
+        "weight": "bert.encoder.layer.5.attention.self.value.MatMul.weight.bin"
+      },
+      "genaiInputNames": {},
+      "inferenceInitializerNames": {
+        "weight": "bert.encoder.layer.5.attention.self.value.MatMul.weight"
+      },
+      "mergedTensorNames": {
+        "weight": "bert.encoder.layer.5.attention.self.value.MatMul.weight"
+      },
+      "mergerOutputNames": {
+        "weight": "merged_weight"
+      },
+      "sha256": {
+        "weight": "71fcbee801f47cad8b13e6f19286d988ac44e778d8616867d8b1b280f6b7e35a"
+      },
+      "shape": [
+        384,
+        384
+      ],
+      "tensorDtypes": {
+        "weight": "float32"
+      },
+      "tensorShapes": {
+        "weight": [
+          384,
+          384
+        ]
+      },
+      "trainingBaseLayerName": "base_model.model.bert.encoder.layer.5.attention.self.value.base_layer",
+      "transposePolicy": "no_transpose"
+    }
+  ],
+  "externalDataLayout": "one_file_per_tensor",
+  "frozenBaseBlob": "frozen_base.onnx.data",
+  "handoffMode": "external_initializer",
+  "mergerModels": {
+    "lora": "merger_lora_fpin_fpout.onnx"
+  },
+  "minReaderVersion": "1.0",
+  "schemaVersion": "1.1"
+}