@camstack/addon-pipeline 1.2.295 → 1.2.296

package/THIRD_PARTY_MODELS.md

@@ -61,6 +61,12 @@ under [Non-commercial models](#non-commercial-models).
   https://github.com/ShiqiYu/libfacedetection.train.
   Trained on WIDER FACE, whose images are CC BY-NC-ND.
 - **AuraFace v1** — fal.ai, Apache-2.0. https://huggingface.co/fal/AuraFace-v1
+- **SigLIP 2** — © Google LLC, Apache-2.0 (`Apache-2.0.txt`).
+  https://huggingface.co/google/siglip2-base-patch16-224 ·
+  https://github.com/google-research/big_vision. CamStack's SigLIP 2 files are
+  modified: the vision input scaling `2x - 1` is baked into the graph, the text
+  graph slices its input to the model's 64 positions, the tokenizer lower-cases,
+  and the towers are converted to ONNX, OpenVINO IR and CoreML FP16.
 - **EasyOCR** — JaidedAI, Apache-2.0. https://github.com/JaidedAI/EasyOCR
 - **U²-Net** — Xuebin Qin et al., Apache-2.0. https://github.com/xuebinqin/U-2-Net
 - **YAMNet** — © Google LLC, Apache-2.0.
@@ -202,6 +208,8 @@ commercial use**:
 | --- | --- | --- | --- | --- | --- |
 | `mobileclip-s1`, `mobileclip-s2` (image) | Apple `ml-mobileclip` | MIT / **Apple Machine Learning Research Model License** | per Apple | Model Derivative: converted to OpenVINO IR and CoreML FP16 | Mirror `clip/mobileclip-s1`, `clip/mobileclip-s2` |
 | `mobileclip-s1-text`, `mobileclip-s2-text` | Apple `ml-mobileclip`, with the CLIP tokenizer | MIT / **Apple Machine Learning Research Model License** | per Apple | Model Derivative: quantised to INT8 ONNX and converted to OpenVINO IR | Mirror `clip/mobileclip-s1`, `clip/mobileclip-s2` |
+| `siglip2-b16-224` (image) | Google SigLIP 2 base patch16-224 (`google/siglip2-base-patch16-224`) | Apache-2.0 / Apache-2.0 | WebLI (Google internal, not released) | input scaling `y = 2x - 1` baked ahead of the patch embedding; converted to OpenVINO IR and CoreML FP16 (an FP32 ONNX is hosted as the verification reference) (`scripts/build-siglip2-model.py`) | Mirror `clip/siglip2` |
+| `siglip2-b16-224-text` | Google SigLIP 2 base patch16-224, with its Gemma tokenizer | Apache-2.0 / Apache-2.0 | WebLI (Google internal, not released) | input sliced to the model's 64 positions in the graph; ONNX weights FP16, OpenVINO IR and CoreML FP16; a `Lowercase` normalizer prepended to the tokenizer | Mirror `clip/siglip2` |
 
 ### Segmentation
 

package/dist/audio-analyzer/index.js

@@ -3,8 +3,8 @@ Object.defineProperties(exports, {
 	[Symbol.toStringTag]: { value: "Module" }
 });
 const require_chunk = require("../chunk-emK7D4bc.js");
-const require_dist = require("../dist-BsVcf5wO.js");
-const require_process_memory = require("../process-memory-CJV29sPf.js");
+const require_dist = require("../dist-8up-f2TX.js");
+const require_process_memory = require("../process-memory-CX_92V_r.js");
 let node_fs = require("node:fs");
 node_fs = require_chunk.__toESM(node_fs);
 let node_path = require("node:path");

package/dist/audio-analyzer/index.mjs

@@ -1,6 +1,6 @@
 import { n as __require } from "../chunk-DnnnRqeS.mjs";
-import { E as HF_BASE_URL, G as audioAnalyzerCapability, Rn as EventCategory, W as audioAnalysisCapability, Wt as resolvePoolMemoryPolicy, _n as hydrateSchema, bt as mapAudioLabelToMacro, en as errMsg, f as DEFAULT_AUDIO_ANALYZER_CONFIG, fn as audioChunkBytesPerSample, gn as expandAudioChunkToF32le, j as PoolMemoryWatchdog, n as AUDIO_BACKEND_CHOICES, sn as BaseAddon, xn as nodePin } from "../dist-CuxSNLKW.mjs";
-import { n as pickNodePlatformArch, t as readProcessMemory } from "../process-memory-D0Nvs9rr.mjs";
+import { E as HF_BASE_URL, G as audioAnalyzerCapability, Gt as resolvePoolMemoryPolicy, Sn as nodePin, W as audioAnalysisCapability, _n as expandAudioChunkToF32le, bt as mapAudioLabelToMacro, cn as BaseAddon, f as DEFAULT_AUDIO_ANALYZER_CONFIG, j as PoolMemoryWatchdog, n as AUDIO_BACKEND_CHOICES, pn as audioChunkBytesPerSample, tn as errMsg, vn as hydrateSchema, zn as EventCategory } from "../dist-CCd0Q3nr.mjs";
+import { n as pickNodePlatformArch, t as readProcessMemory } from "../process-memory-DFC_O5zE.mjs";
 import * as fs from "node:fs";
 import * as path$1 from "node:path";
 import { downloadFile } from "@camstack/system/addon-utils";

package/dist/{default-detection-model-CAUBVbgK.mjs → default-detection-model-Co578D8C.mjs}

@@ -1,4 +1,4 @@
-import { K as audioKindId, d as COCO_TO_MACRO, ht as hfModelUrl, r as AUDIO_MACRO_LABELS, u as COCO_80_LABELS } from "./dist-CuxSNLKW.mjs";
+import { K as audioKindId, d as COCO_TO_MACRO, ht as hfModelUrl, r as AUDIO_MACRO_LABELS, u as COCO_80_LABELS } from "./dist-CCd0Q3nr.mjs";
 import { randomUUID } from "node:crypto";
 //#region src/detection-pipeline/pipeline/landmark-precision-gate.ts
 /**
@@ -1184,55 +1184,84 @@ var INSTANCE_SEGMENTATION_MODELS = [
 		}
 	}
 ];
-var CLIP_EMBEDDING_MODELS = [{
-	id: "mobileclip-s1",
-	name: "MobileCLIP S1",
-	description: "MobileCLIP S1 — Apple balanced CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
-	inputSize: {
-		width: 256,
-		height: 256
+var CLIP_EMBEDDING_MODELS = [
+	{
+		id: "mobileclip-s1",
+		name: "MobileCLIP S1",
+		description: "MobileCLIP S1 — Apple balanced CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 256,
+			height: 256
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		formats: {
+			openvino: ovFormat(hf("clip/mobileclip-s1/openvino/camstack-mobileclip-s1-vision.xml"), 55),
+			coreml: {
+				url: hf("clip/mobileclip-s1/coreml/camstack-mobileclip-s1-vision.mlpackage"),
+				sizeMB: 65,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
+		}
 	},
-	labels: [{
-		id: "embedding",
-		name: "CLIP Embedding"
-	}],
-	preprocessMode: "resize",
-	inputNormalization: "none",
-	formats: {
-		openvino: ovFormat(hf("clip/mobileclip-s1/openvino/camstack-mobileclip-s1-vision.xml"), 55),
-		coreml: {
-			url: hf("clip/mobileclip-s1/coreml/camstack-mobileclip-s1-vision.mlpackage"),
-			sizeMB: 65,
-			isDirectory: true,
-			files: [...MLPACKAGE_FILES],
-			runtimes: ["python"]
+	{
+		id: "mobileclip-s2",
+		name: "MobileCLIP S2",
+		description: "MobileCLIP S2 — Apple high-accuracy CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 256,
+			height: 256
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		formats: {
+			openvino: ovFormat(hf("clip/mobileclip-s2/openvino/camstack-mobileclip-s2-vision.xml"), 90),
+			coreml: {
+				url: hf("clip/mobileclip-s2/coreml/camstack-mobileclip-s2-vision.mlpackage"),
+				sizeMB: 110,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
 		}
-	}
-}, {
-	id: "mobileclip-s2",
-	name: "MobileCLIP S2",
-	description: "MobileCLIP S2 — Apple high-accuracy CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
-	inputSize: {
-		width: 256,
-		height: 256
 	},
-	labels: [{
-		id: "embedding",
-		name: "CLIP Embedding"
-	}],
-	preprocessMode: "resize",
-	inputNormalization: "none",
-	formats: {
-		openvino: ovFormat(hf("clip/mobileclip-s2/openvino/camstack-mobileclip-s2-vision.xml"), 90),
-		coreml: {
-			url: hf("clip/mobileclip-s2/coreml/camstack-mobileclip-s2-vision.mlpackage"),
-			sizeMB: 110,
-			isDirectory: true,
-			files: [...MLPACKAGE_FILES],
-			runtimes: ["python"]
+	{
+		id: "siglip2-b16-224",
+		name: "SigLIP2 B/16",
+		description: "Google SigLIP2 base, patch 16, 224×224 — Apache-2.0 CLIP vision encoder, 768-dim (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 224,
+			height: 224
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		license: "Apache-2.0",
+		formats: {
+			openvino: ovFormat(hf("clip/siglip2/openvino/camstack-siglip2-b16-224-vision.xml"), 186),
+			coreml: {
+				url: hf("clip/siglip2/coreml/camstack-siglip2-b16-224-vision.mlpackage"),
+				sizeMB: 185,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
 		}
 	}
-}];
+];
 var AUDIO_CLASSIFIER_MODELS = [{
 	id: "yamnet-onnx",
 	name: "YAMNet",
@@ -1671,7 +1700,7 @@ var STEP_FACE_EMBEDDING = new FaceEmbeddingStep({
 	outputClasses: ["identity"],
 	labelTier: 2,
 	models: [...FACE_EMBEDDING_MODELS],
-	defaultModelId: "arcface-r100",
+	defaultModelId: "auraface-r100",
 	modelScope: "cluster",
 	defaultConfidence: 0,
 	cadence: {

package/dist/{default-detection-model-wNEISVA9.js → default-detection-model-D1daTtqT.js}

@@ -1,4 +1,4 @@
-const require_dist = require("./dist-BsVcf5wO.js");
+const require_dist = require("./dist-8up-f2TX.js");
 let node_crypto = require("node:crypto");
 //#region src/detection-pipeline/pipeline/landmark-precision-gate.ts
 /**
@@ -1184,55 +1184,84 @@ var INSTANCE_SEGMENTATION_MODELS = [
 		}
 	}
 ];
-var CLIP_EMBEDDING_MODELS = [{
-	id: "mobileclip-s1",
-	name: "MobileCLIP S1",
-	description: "MobileCLIP S1 — Apple balanced CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
-	inputSize: {
-		width: 256,
-		height: 256
+var CLIP_EMBEDDING_MODELS = [
+	{
+		id: "mobileclip-s1",
+		name: "MobileCLIP S1",
+		description: "MobileCLIP S1 — Apple balanced CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 256,
+			height: 256
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		formats: {
+			openvino: ovFormat(hf("clip/mobileclip-s1/openvino/camstack-mobileclip-s1-vision.xml"), 55),
+			coreml: {
+				url: hf("clip/mobileclip-s1/coreml/camstack-mobileclip-s1-vision.mlpackage"),
+				sizeMB: 65,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
+		}
 	},
-	labels: [{
-		id: "embedding",
-		name: "CLIP Embedding"
-	}],
-	preprocessMode: "resize",
-	inputNormalization: "none",
-	formats: {
-		openvino: ovFormat(hf("clip/mobileclip-s1/openvino/camstack-mobileclip-s1-vision.xml"), 55),
-		coreml: {
-			url: hf("clip/mobileclip-s1/coreml/camstack-mobileclip-s1-vision.mlpackage"),
-			sizeMB: 65,
-			isDirectory: true,
-			files: [...MLPACKAGE_FILES],
-			runtimes: ["python"]
+	{
+		id: "mobileclip-s2",
+		name: "MobileCLIP S2",
+		description: "MobileCLIP S2 — Apple high-accuracy CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 256,
+			height: 256
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		formats: {
+			openvino: ovFormat(hf("clip/mobileclip-s2/openvino/camstack-mobileclip-s2-vision.xml"), 90),
+			coreml: {
+				url: hf("clip/mobileclip-s2/coreml/camstack-mobileclip-s2-vision.mlpackage"),
+				sizeMB: 110,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
 		}
-	}
-}, {
-	id: "mobileclip-s2",
-	name: "MobileCLIP S2",
-	description: "MobileCLIP S2 — Apple high-accuracy CLIP vision encoder, 512-dim, 256×256 (fp16 OpenVINO/CoreML)",
-	inputSize: {
-		width: 256,
-		height: 256
 	},
-	labels: [{
-		id: "embedding",
-		name: "CLIP Embedding"
-	}],
-	preprocessMode: "resize",
-	inputNormalization: "none",
-	formats: {
-		openvino: ovFormat(hf("clip/mobileclip-s2/openvino/camstack-mobileclip-s2-vision.xml"), 90),
-		coreml: {
-			url: hf("clip/mobileclip-s2/coreml/camstack-mobileclip-s2-vision.mlpackage"),
-			sizeMB: 110,
-			isDirectory: true,
-			files: [...MLPACKAGE_FILES],
-			runtimes: ["python"]
+	{
+		id: "siglip2-b16-224",
+		name: "SigLIP2 B/16",
+		description: "Google SigLIP2 base, patch 16, 224×224 — Apache-2.0 CLIP vision encoder, 768-dim (fp16 OpenVINO/CoreML)",
+		inputSize: {
+			width: 224,
+			height: 224
+		},
+		labels: [{
+			id: "embedding",
+			name: "CLIP Embedding"
+		}],
+		preprocessMode: "resize",
+		inputNormalization: "none",
+		license: "Apache-2.0",
+		formats: {
+			openvino: ovFormat(hf("clip/siglip2/openvino/camstack-siglip2-b16-224-vision.xml"), 186),
+			coreml: {
+				url: hf("clip/siglip2/coreml/camstack-siglip2-b16-224-vision.mlpackage"),
+				sizeMB: 185,
+				isDirectory: true,
+				files: [...MLPACKAGE_FILES],
+				runtimes: ["python"]
+			}
 		}
 	}
-}];
+];
 var AUDIO_CLASSIFIER_MODELS = [{
 	id: "yamnet-onnx",
 	name: "YAMNet",
@@ -1671,7 +1700,7 @@ var STEP_FACE_EMBEDDING = new FaceEmbeddingStep({
 	outputClasses: ["identity"],
 	labelTier: 2,
 	models: [...FACE_EMBEDDING_MODELS],
-	defaultModelId: "arcface-r100",
+	defaultModelId: "auraface-r100",
 	modelScope: "cluster",
 	defaultConfidence: 0,
 	cadence: {