diff --git a/.cspell-wordlist.txt b/.cspell-wordlist.txt index 55fa014084..006f5be69f 100644 --- a/.cspell-wordlist.txt +++ b/.cspell-wordlist.txt @@ -298,3 +298,11 @@ deis Coeff Multistep multistep +Picabia +Picabia's +atrous +Convolutional +Roboflow +Ultralytics + + diff --git a/packages/react-native-executorch/src/models.ts b/packages/react-native-executorch/src/models.ts index fc8e7a5e7d..31b696ad74 100644 --- a/packages/react-native-executorch/src/models.ts +++ b/packages/react-native-executorch/src/models.ts @@ -754,7 +754,17 @@ const ALL_MINILM_L6_V2_TOKENIZER = `${BASE_URL}-all-MiniLM-L6-v2/${VERSION_TAG}/ * @category Utils */ export const models = { + /** + * Image classification models that categorize input images into pre-defined + * classes. + */ classification: { + /** + * EfficientNetV2-S image classification model pre-trained on ImageNet-1k + * (1000 categories, see {@link IMAGENET1K_LABELS}). + * Compact and efficient architecture providing high accuracy for + * general-purpose image classification. + */ EFFICIENTNET_V2_S: { ...EFFICIENTNET_V2_S_XNNPACK_INT8, XNNPACK_INT8: EFFICIENTNET_V2_S_XNNPACK_INT8, @@ -762,7 +772,16 @@ export const models = { COREML_FP16: EFFICIENTNET_V2_S_COREML_FP16, }, }, + + /** + * Artistic style transfer models that re-style input images according to + * artwork patterns. + */ styleTransfer: { + /** + * Fast neural style transfer model generating a vibrant, artistic "Candy" + * style effect. + */ CANDY: { ...STYLE_TRANSFER_CANDY_XNNPACK_INT8, XNNPACK_FP32: STYLE_TRANSFER_CANDY_XNNPACK_FP32, @@ -770,6 +789,10 @@ export const models = { COREML_FP16: STYLE_TRANSFER_CANDY_COREML_FP16, COREML_FP32: STYLE_TRANSFER_CANDY_COREML_FP32, }, + /** + * Fast neural style transfer model applying a classic tile mosaic + * artistic pattern. + */ MOSAIC: { ...STYLE_TRANSFER_MOSAIC_XNNPACK_INT8, XNNPACK_FP32: STYLE_TRANSFER_MOSAIC_XNNPACK_FP32, @@ -777,6 +800,10 @@ export const models = { COREML_FP16: STYLE_TRANSFER_MOSAIC_COREML_FP16, COREML_FP32: STYLE_TRANSFER_MOSAIC_COREML_FP32, }, + /** + * Fast neural style transfer model applying a painterly "Rain Princess" + * oil painting aesthetic. + */ RAIN_PRINCESS: { ...STYLE_TRANSFER_RAIN_PRINCESS_XNNPACK_INT8, XNNPACK_FP32: STYLE_TRANSFER_RAIN_PRINCESS_XNNPACK_FP32, @@ -784,6 +811,10 @@ export const models = { COREML_FP16: STYLE_TRANSFER_RAIN_PRINCESS_COREML_FP16, COREML_FP32: STYLE_TRANSFER_RAIN_PRINCESS_COREML_FP32, }, + /** + * Fast neural style transfer model applying Francis Picabia's "Udnie" + * abstract art style. + */ UDNIE: { ...STYLE_TRANSFER_UDNIE_XNNPACK_INT8, XNNPACK_FP32: STYLE_TRANSFER_UDNIE_XNNPACK_FP32, @@ -792,80 +823,158 @@ export const models = { COREML_FP32: STYLE_TRANSFER_UDNIE_COREML_FP32, }, }, + + /** + * Semantic segmentation models that classify each pixel into target object + * or background classes. + */ semanticSegmentation: { + /** + * Lightweight portrait selfie segmentation model for real-time person vs + * background separation. Categorizes pixels into `background` and + * `person`. Ideal for background blur and replacement effects. + */ SELFIE_SEGMENTATION: { ...SELFIE_SEGMENTATION_XNNPACK_FP32, XNNPACK_FP32: SELFIE_SEGMENTATION_XNNPACK_FP32, }, + /** + * Lite R-ASPP semantic segmentation model with MobileNetV3-Large backbone + * (21 classes, see {@link PASCAL_VOC_LABELS}). Optimized for low-latency, + * real-time pixel-level segmentation on mobile devices. + */ LRASPP_MOBILENET_V3_LARGE: { ...LRASPP_MOBILENET_V3_LARGE_XNNPACK_INT8, XNNPACK_FP32: LRASPP_MOBILENET_V3_LARGE_XNNPACK_FP32, XNNPACK_INT8: LRASPP_MOBILENET_V3_LARGE_XNNPACK_INT8, }, + /** + * DeepLabV3 semantic segmentation model with ResNet-50 backbone + * (21 classes, see {@link PASCAL_VOC_LABELS}). High-accuracy segmentation + * utilizing atrous spatial pyramid pooling. + */ DEEPLAB_V3_RESNET50: { ...DEEPLAB_V3_RESNET50_XNNPACK_INT8, XNNPACK_FP32: DEEPLAB_V3_RESNET50_XNNPACK_FP32, XNNPACK_INT8: DEEPLAB_V3_RESNET50_XNNPACK_INT8, }, + /** + * DeepLabV3 semantic segmentation model with ResNet-101 backbone + * (21 classes, see {@link PASCAL_VOC_LABELS}). High-capacity backbone for + * maximum segmentation detail and boundary accuracy. + */ DEEPLAB_V3_RESNET101: { ...DEEPLAB_V3_RESNET101_XNNPACK_INT8, XNNPACK_FP32: DEEPLAB_V3_RESNET101_XNNPACK_FP32, XNNPACK_INT8: DEEPLAB_V3_RESNET101_XNNPACK_INT8, }, + /** + * DeepLabV3 semantic segmentation model with MobileNetV3-Large backbone + * (21 classes, see {@link PASCAL_VOC_LABELS}). Combines DeepLabV3 feature + * extraction quality with a lightweight mobile backbone. + */ DEEPLAB_V3_MOBILENET_V3_LARGE: { ...DEEPLAB_V3_MOBILENET_V3_LARGE_XNNPACK_INT8, XNNPACK_FP32: DEEPLAB_V3_MOBILENET_V3_LARGE_XNNPACK_FP32, XNNPACK_INT8: DEEPLAB_V3_MOBILENET_V3_LARGE_XNNPACK_INT8, }, + /** + * Fully Convolutional Network (FCN) semantic segmentation model with + * ResNet-50 backbone (21 classes, see {@link PASCAL_VOC_LABELS}). + */ FCN_RESNET50: { ...FCN_RESNET50_XNNPACK_INT8, XNNPACK_FP32: FCN_RESNET50_XNNPACK_FP32, XNNPACK_INT8: FCN_RESNET50_XNNPACK_INT8, }, + /** + * Fully Convolutional Network (FCN) semantic segmentation model with + * ResNet-101 backbone (21 classes, see {@link PASCAL_VOC_LABELS}). + */ FCN_RESNET101: { ...FCN_RESNET101_XNNPACK_INT8, XNNPACK_FP32: FCN_RESNET101_XNNPACK_FP32, XNNPACK_INT8: FCN_RESNET101_XNNPACK_INT8, }, }, + + /** + * Object detection models that identify object locations and bounding boxes. + */ objectDetection: { + /** + * SSDLite object detector with MobileNetV3-Large backbone trained on COCO + * (see {@link COCO_CLASSES}) at 320x320 resolution. Fast, lightweight + * detector suited for real-time mobile applications. + */ SSDLITE320_MOBILENET_V3_LARGE: { ...SSDLITE320_MOBILENET_V3_LARGE_XNNPACK_FP32, XNNPACK_FP32: SSDLITE320_MOBILENET_V3_LARGE_XNNPACK_FP32, COREML_FP16: SSDLITE320_MOBILENET_V3_LARGE_COREML_FP16, COREML_FP32: SSDLITE320_MOBILENET_V3_LARGE_COREML_FP32, }, + /** + * RF-DETR (Roboflow Detection Transformer) Nano variant trained on COCO + * (see {@link COCO_CLASSES}). Modern end-to-end DINOv2-based transformer + * object detector. + */ RFDETR_NANO: { ...RFDETR_NANO_DETECTOR_XNNPACK_FP32, XNNPACK_FP32: RFDETR_NANO_DETECTOR_XNNPACK_FP32, COREML_INT8: RFDETR_NANO_DETECTOR_COREML_INT8, }, + /** + * Ultralytics YOLO26 real-time object detection models trained on COCO + * (80 classes, see {@link COCO_CLASSES_YOLO}). Available across multiple + * scale sizes (NANO, SMALL, MEDIUM, LARGE, XLARGE) and resolutions + * (384x384, 512x512, 640x640). + */ YOLO26: { ...YOLO26_NANO_384_XNNPACK_FP32, + /** + * Nano scale YOLO26 object detection model. High speed, + * ultra low latency. + */ NANO: { ...YOLO26_NANO_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_NANO_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_NANO_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_NANO_640_XNNPACK_FP32 }, }, + /** + * Small scale YOLO26 object detection model. Balanced + * latency and accuracy. + */ SMALL: { ...YOLO26_SMALL_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_SMALL_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_SMALL_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_SMALL_640_XNNPACK_FP32 }, }, + /** + * Medium scale YOLO26 object detection model. Higher + * precision for complex scenes. + */ MEDIUM: { ...YOLO26_MEDIUM_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_MEDIUM_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_MEDIUM_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_MEDIUM_640_XNNPACK_FP32 }, }, + /** + * Large scale YOLO26 object detection model. High accuracy + * model variant. + */ LARGE: { ...YOLO26_LARGE_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_LARGE_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_LARGE_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_LARGE_640_XNNPACK_FP32 }, }, + /** + * Extra Large scale YOLO26 object detection model. + * Maximum detection performance. + */ XLARGE: { ...YOLO26_XLARGE_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_XLARGE_384_XNNPACK_FP32 }, @@ -874,17 +983,35 @@ export const models = { }, }, }, + + /** + * Keypoint and pose detection models that estimate facial landmarks or human + * body skeletal keypoints. + */ keypointDetection: { + /** + * MediaPipe BlazeFace lightweight face detection and 6-point facial landmark + * locator (eyes, nose, mouth, ears, see {@link BLAZEFACE_LANDMARKS}). + */ BLAZEFACE: { ...BLAZEFACE_XNNPACK_FP32, XNNPACK_FP32: BLAZEFACE_XNNPACK_FP32, }, + /** + * YOLO26 human pose estimation model predicting 17 COCO body keypoints + * (see {@link COCO_LANDMARKS}). Available across 384x384, 512x512, and + * 640x640 resolutions. + */ YOLO26_POSE: { ...YOLO26_POSE_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_POSE_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_POSE_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_POSE_640_XNNPACK_FP32 }, }, + /** + * RF-DETR (Roboflow Detection Transformer) pose keypoint detector + * predicting 17 COCO body keypoints (see {@link COCO_LANDMARKS}). + */ RFDETR_KEYPOINT: { ...RFDETR_KEYPOINT_XNNPACK_FP32, XNNPACK_FP32: RFDETR_KEYPOINT_XNNPACK_FP32, @@ -892,14 +1019,31 @@ export const models = { MLX_FP32: RFDETR_KEYPOINT_MLX_FP32, }, }, + + /** + * Instance segmentation models predicting both bounding boxes and fine-grained + * pixel masks per object instance. + */ instanceSegmentation: { + /** + * Fast Segment Anything Model (FastSAM) for promptable or global object + * instance mask segmentation. Available in Small (S) and Extra Large (X) + * variants. + */ FASTSAM: { + /** + * FastSAM Small - lightweight instance segmenter for + * mobile. + */ S: { ...FASTSAM_S_XNNPACK_FP32, XNNPACK_FP32: FASTSAM_S_XNNPACK_FP32, COREML_FP32: FASTSAM_S_COREML_FP32, COREML_FP16: FASTSAM_S_COREML_FP16, }, + /** + * FastSAM Extra Large - high-accuracy instance segmenter. + */ X: { ...FASTSAM_X_XNNPACK_FP32, XNNPACK_FP32: FASTSAM_X_XNNPACK_FP32, @@ -907,37 +1051,56 @@ export const models = { COREML_FP16: FASTSAM_X_COREML_FP16, }, }, + /** + * RF-DETR (Roboflow Detection Transformer) Nano instance segmentation + * model predicting COCO class masks and bounding boxes + * (see {@link COCO_CLASSES}). + */ RFDETR_NANO: { ...RFDETR_NANO_SEG_COREML_INT8, COREML_INT8: RFDETR_NANO_SEG_COREML_INT8, XNNPACK_FP32: RFDETR_NANO_SEG_XNNPACK_FP32, }, + /** + * YOLO26 instance segmentation models predicting COCO class instance masks + * and bounding boxes (see {@link COCO_CLASSES_YOLO}). Available across + * multiple sizes (NANO, SMALL, MEDIUM, LARGE, XLARGE) and resolutions + * (384x384, 512x512, 640x640). + */ YOLO26: { ...YOLO26_NANO_SEG_384_XNNPACK_FP32, + /** + * Nano scale YOLO26 instance segmentation model. High + * speed. + */ NANO: { ...YOLO26_NANO_SEG_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_NANO_SEG_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_NANO_SEG_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_NANO_SEG_640_XNNPACK_FP32 }, }, + /** Small scale YOLO26 instance segmentation model. */ SMALL: { ...YOLO26_SMALL_SEG_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_SMALL_SEG_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_SMALL_SEG_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_SMALL_SEG_640_XNNPACK_FP32 }, }, + /** Medium scale YOLO26 instance segmentation model. */ MEDIUM: { ...YOLO26_MEDIUM_SEG_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_MEDIUM_SEG_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_MEDIUM_SEG_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_MEDIUM_SEG_640_XNNPACK_FP32 }, }, + /** Large scale YOLO26 instance segmentation model. */ LARGE: { ...YOLO26_LARGE_SEG_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_LARGE_SEG_384_XNNPACK_FP32 }, SIZE_512: { XNNPACK_FP32: YOLO26_LARGE_SEG_512_XNNPACK_FP32 }, SIZE_640: { XNNPACK_FP32: YOLO26_LARGE_SEG_640_XNNPACK_FP32 }, }, + /** Extra Large scale YOLO26 instance segmentation model. */ XLARGE: { ...YOLO26_XLARGE_SEG_384_XNNPACK_FP32, SIZE_384: { XNNPACK_FP32: YOLO26_XLARGE_SEG_384_XNNPACK_FP32 }, @@ -946,47 +1109,89 @@ export const models = { }, }, }, + + /** + * Voice Activity Detection (VAD) models detecting speech vs non-speech + * intervals in real-time audio streams. + */ voiceActivityDetection: { + /** + * Feedforward Sequential Memory Network (FSMN) Voice Activity Detection + * model. Extremely lightweight model evaluating continuous speech + * probability chunks for live mic streaming and STT preprocessing. + */ FSMN_VAD: { ...FSMN_VAD_XNNPACK_FP32, XNNPACK_FP32: FSMN_VAD_XNNPACK_FP32, }, }, + + /** + * Automatic Speech Recognition (ASR) / Speech-to-Text models. + */ speechToText: { + /** + * OpenAI Whisper automatic speech recognition model family with integrated + * Voice Activity Detection. Includes multilingual and English-only (`EN`) + * variants across model sizes (`TINY`, `BASE`, `SMALL`). + */ WHISPER: { - ...WHISPER_TINY_XNNPACK_FP32, + /** + * Multilingual Whisper Tiny model. Supporting 99+ + * languages. High speed speech recognition. + */ TINY: { ...WHISPER_TINY_XNNPACK_FP32, XNNPACK_FP32: WHISPER_TINY_XNNPACK_FP32, COREML_FP16: WHISPER_TINY_COREML_FP16, MLX_BF16: WHISPER_TINY_MLX_BF16, }, + /** + * Multilingual Whisper Base model. Higher accuracy across + * supported languages. + */ BASE: { ...WHISPER_BASE_XNNPACK_FP32, XNNPACK_FP32: WHISPER_BASE_XNNPACK_FP32, COREML_FP16: WHISPER_BASE_COREML_FP16, MLX_BF16: WHISPER_BASE_MLX_BF16, }, + /** + * Multilingual Whisper Small model. Best accuracy for + * complex multi-language audio. + */ SMALL: { ...WHISPER_SMALL_XNNPACK_FP32, XNNPACK_FP32: WHISPER_SMALL_XNNPACK_FP32, COREML_FP16: WHISPER_SMALL_COREML_FP16, MLX_BF16: WHISPER_SMALL_MLX_BF16, }, + /** English-only optimized Whisper models (`TINY`, `BASE`, `SMALL`). */ EN: { - ...WHISPER_TINY_EN_XNNPACK_FP32, + /** + * English-only Whisper Tiny model. Fast and compact for + * English STT. + */ TINY: { ...WHISPER_TINY_EN_XNNPACK_FP32, XNNPACK_FP32: WHISPER_TINY_EN_XNNPACK_FP32, COREML_FP16: WHISPER_TINY_EN_COREML_FP16, MLX_BF16: WHISPER_TINY_EN_MLX_BF16, }, + /** + * English-only Whisper Base model. High accuracy + * English speech recognition. + */ BASE: { ...WHISPER_BASE_EN_XNNPACK_FP32, XNNPACK_FP32: WHISPER_BASE_EN_XNNPACK_FP32, COREML_FP16: WHISPER_BASE_EN_COREML_FP16, MLX_BF16: WHISPER_BASE_EN_MLX_BF16, }, + /** + * English-only Whisper Small model. Superior accuracy + * for English transcription. + */ SMALL: { ...WHISPER_SMALL_EN_XNNPACK_FP32, XNNPACK_FP32: WHISPER_SMALL_EN_XNNPACK_FP32, @@ -996,41 +1201,83 @@ export const models = { }, }, }, + + /** + * Standalone text tokenizers for preprocessing strings into token ID arrays. + */ tokenizer: { + /** WordPiece tokenizer URL for the `all-MiniLM-L6-v2` embedding model. */ ALL_MINILM_L6_V2: ALL_MINILM_L6_V2_TOKENIZER, }, + + /** + * Text embedding models mapping sentences and documents into dense vector + * representations for semantic search and RAG. + */ textEmbeddings: { + /** + * Compact 384-dimensional sentence transformer mapping text to a dense + * vector space. Optimized for fast, general-purpose semantic search, + * sentence similarity, and clustering. + */ ALL_MINILM_L6_V2: { ...ALL_MINILM_L6_V2_EMBEDDINGS, XNNPACK_FP32: ALL_MINILM_L6_V2_EMBEDDINGS, }, + /** + * High-quality 768-dimensional sentence transformer model based on MPNet. + * Provides higher quality semantic embeddings compared to MiniLM. + */ ALL_MPNET_BASE_V2: { ...ALL_MPNET_BASE_V2_EMBEDDINGS, XNNPACK_FP32: ALL_MPNET_BASE_V2_EMBEDDINGS, }, + /** + * 384-dimensional sentence transformer fine-tuned specifically for + * semantic QA matching using cosine similarity. + */ MULTI_QA_MINILM_L6_COS_V1: { ...MULTI_QA_MINILM_L6_COS_V1_EMBEDDINGS, XNNPACK_FP32: MULTI_QA_MINILM_L6_COS_V1_EMBEDDINGS, }, + /** + * 768-dimensional sentence transformer fine-tuned specifically for + * question-answering matching using dot product distance. + */ MULTI_QA_MPNET_BASE_DOT_V1: { ...MULTI_QA_MPNET_BASE_DOT_V1_EMBEDDINGS, XNNPACK_FP32: MULTI_QA_MPNET_BASE_DOT_V1_EMBEDDINGS, }, + /** + * 384-dimensional sentence transformer supporting 50+ languages for + * cross-lingual semantic similarity. + */ PARAPHRASE_MULTILINGUAL_MINILM_L12_V2: { ...PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_EMBEDDINGS, XNNPACK_8DA4W: PARAPHRASE_MULTILINGUAL_MINILM_L12_V2_EMBEDDINGS, }, + /** + * Multilingual sentence transformer supporting 50+ languages, based on + * distilled Universal Sentence Encoder (512-dim output). + */ DISTILUSE_BASE_MULTILINGUAL_CASED_V2: { ...DISTILUSE_BASE_MULTILINGUAL_CASED_V2_EMBEDDINGS, XNNPACK_8DA4W: DISTILUSE_BASE_MULTILINGUAL_CASED_V2_EMBEDDINGS, }, + /** + * CLIP text encoder (ViT-B/32) mapping text queries into a + * 512-dimensional joint text-image embedding space. Used in combination + * with `imageEmbeddings.CLIP_VIT_BASE_PATCH32` for zero-shot text-to-image + * search. + */ CLIP_VIT_BASE_PATCH32_TEXT: { ...CLIP_VIT_BASE_PATCH32_TEXT_EMBEDDINGS, XNNPACK_FP32: CLIP_VIT_BASE_PATCH32_TEXT_EMBEDDINGS, }, /** - * Asymmetric retrieval model: prompt queries with `query: ` (the default) - * and passages with `document: ` via `embed(text, 'document: ')`. + * Liquid AI LFM 2.5 350M parameter embedding model for asymmetric search + * and retrieval tasks. Prompts queries with `query: ` (the default) and + * passages with `document: ` via `embed(text, 'document: ')`. */ LFM2_5_EMBEDDING_350M: { ...LFM2_5_EMBEDDING_350M_EMBEDDINGS, @@ -1038,14 +1285,32 @@ export const models = { MLX_INT4: LFM2_5_EMBEDDING_350M_MLX_INT4, }, }, + + /** + * Image feature extraction and vision embedding models. + */ imageEmbeddings: { + /** + * CLIP vision encoder (ViT-B/32) mapping images into a 512-dimensional + * shared text-image space. Used for zero-shot visual classification and + * cross-modal image search. + */ CLIP_VIT_BASE_PATCH32: { ...CLIP_VIT_BASE_PATCH32_IMAGE_XNNPACK_FP32, XNNPACK_FP32: CLIP_VIT_BASE_PATCH32_IMAGE_XNNPACK_FP32, XNNPACK_INT8: CLIP_VIT_BASE_PATCH32_IMAGE_XNNPACK_INT8, }, }, + + /** + * Generative text-to-image synthesis models. + */ textToImage: { + /** + * Ultra-fast SDXS (Stable Diffusion eXtreme Speed) 512x512 text-to-image + * generation model based on DreamShaper. Generates high-quality images + * from text prompts in real time. + */ SDXS_512_DREAMSHAPER: { ...SDXS_512_DREAMSHAPER_XNNPACK_FP32, XNNPACK_FP32: SDXS_512_DREAMSHAPER_XNNPACK_FP32,