manifest.json

[
	{
		"name": "./arc_fault_detection/st_conv/ST_pretrainedmodel_custom_dataset/afd_test_bench_dataset/st_conv_time_4channels_512/st_conv_time_4channels_512_int8.tflite",
		"description": "st_conv model is a compact per‑channel CNN for arc fault detection (normal vs arc). Trained on ST proprietary AFD test bench dataset",
		"usecase": "arc_fault_detection",
		"dataset_used": "afd_test_bench_dataset",
		"readme": "./arc_fault_detection/st_conv/README.md",
		"visible": true
	},
	{
		"name": "./arc_fault_detection/st_dense/ST_pretrainedmodel_custom_dataset/afd_test_bench_dataset/st_dense_freq_8channels_512/st_dense_freq_8channels_512_int8.tflite",
		"description": " st_dense model is a lightweight fully connected neural network designed for arc fault detection (normal vs arc). Trained on ST proprietary AFD test bench dataset",
		"usecase": "arc_fault_detection",
		"dataset_used": "afd_test_bench_dataset",
		"readme": "./arc_fault_detection/st_dense/README.md",
		"visible": true
	},
	{
		"name": "./audio_event_detection/miniresnetv1/ST_pretrainedmodel_public_dataset/esc10/miniresnetv1_s1_64x50_tl/miniresnetv1_s1_64x50_tl_int8.tflite",
		"description": "MiniResNet v1 trained on ESC-10 with 10 classes and input mel-spectrogram size 64x50",
		"usecase": "audio_event_detection",
		"dataset_used": "esc10",
		"readme": "./audio_event_detection/miniresnet/README.md",
		"visible": true
	},
	{
		"name": "./audio_event_detection/yamnet/ST_pretrainedmodel_public_dataset/esc10/yamnet_e1024_64x96_tl/yamnet_e1024_64x96_tl_qdq_int8.onnx",
		"description": "Yamnet trained on ESC-10 with 10 classes, embedding size 1024 and input mel-spectrogram size 64x96",
		"usecase": "audio_event_detection",
		"dataset_used": "esc10",
		"readme": "./audio_event_detection/yamnet/README.md",
		"visible": true
	},
	{
		"name": "./audio_event_detection/yamnet/ST_pretrainedmodel_public_dataset/fsd50k/yamnet_e256_64x96_tl/without_unknown_class/yamnet_e256_64x96_tl_int8.tflite",
		"description": "Yamnet trained on FSD50K with 5 classes, embedding size 256 and input mel-spectrogram size 64x96, without unknown class",
		"usecase": "audio_event_detection",
		"dataset_used": "fsd50k",
		"readme": "./audio_event_detection/yamnet/README.md",
		"visible": true
	},
	{
		"name": "./depth_estimation/fastdepth/Public_pretrainedmodel_public_dataset/nyu_depthv2/fastdepth_224/fastdepth_224_int8.tflite",
		"description": "FastDepth model trained on NYU Depth v2 dataset (input 224x224 rgb image)",
		"usecase": "depth_estimation",
		"dataset_used": "nyu_depthv2",
		"readme": "./depth_estimation/fastdepth/README.md",
		"visible": true
	},
	{
		"name": "./depth_estimation/fastdepth/Public_pretrainedmodel_public_dataset/nyu_depthv2/fastdepth_320/fastdepth_320_int8.tflite",
		"description": "FastDepth model trained on NYU Depth v2 dataset (input 320x320 rgb image)",
		"usecase": "depth_estimation",
		"dataset_used": "nyu_depthv2",
		"readme": "./depth_estimation/fastdepth/README.md",
		"visible": true
	},
	{
		"name": "./face_detection/yunet/Public_pretrainedmodel_public_dataset/widerface/yunetn_320/yunetn_320_qdq_int8.onnx",
		"description": "YuNet model trained on WiderFace dataset (input size 320x320 rgb image)",
		"usecase": "face_detection",
		"dataset_used": "widerface",
		"readme": "./face_detection/yunet/README.md",
		"visible": true
	},
	{
		"name": "./face_detection/facedetect_front/Public_pretrainedmodel_public_dataset/widerface/blazeface_front_128/blazeface_front_128_int8.tflite",
		"description": "BlazeFace front model trained on WiderFace dataset (input size 128x128 rgb image)",
		"usecase": "face_detection",
		"dataset_used": "widerface",
		"readme": "./face_detection/facedetect_front/README.md",
		"visible": true
	},
	{
	"name": "./hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L5CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras",
	"description": "st_cnn2d_handposture_postures model trained on ST_VL53L5CX_handposture_dataset with 8 classes",
	"usecase": "hand_posture",
	"dataset_used": "ST_VL53L5CX_handposture_dataset",
	"readme": "./hand_posture/st_cnn2d_handposture/README.md",
		"visible": true
	},
	{
	"name": "./hand_posture/st_cnn2d_handposture/ST_pretrainedmodel_custom_dataset/ST_VL53L8CX_handposture_dataset/st_cnn2d_handposture_8classes/st_cnn2d_handposture_8classes.keras",
	"description": "st_cnn2d_handposture_postures model trained on ST_VL53L8CX_handposture_dataset with 8 classes",
	"usecase": "hand_posture",
	"dataset_used": "ST_VL53L8CX_handposture_dataset",
	"readme": "./hand_posture/st_cnn2d_handposture/README.md",
		"visible": true
	},
	{
		"name": "./human_activity_recognition/st_gmp/ST_pretrainedmodel_public_dataset/WISDM/st_gmp_wl_48/st_gmp_wl_48.keras",
		"description": "GMP model trained on WISDM with 4 classes and 48x3 input shape (accelerometer data with 48 time steps and 3 axes)",
		"usecase": "human_activity_recognition",
		"dataset_used": "WISDM",
		"readme": "./human_activity_recognition/st_gmp/README.md",
		"visible": true
	},
	{
		"name": "./human_activity_recognition/st_ign/ST_pretrainedmodel_custom_dataset/mobility_v1/st_ign_wl_48/st_ign_wl_48.keras",
		"description": "IGN model trained on mobility_v1 (an ST proprietary dataset) with 4 classes and 48x3 input shape (accelerometer data with 48 time steps and 3 axes)",
		"usecase": "human_activity_recognition",
		"dataset_used": "mobility_v1",
		"readme": "./human_activity_recognition/st_ign/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/efficientnet/ST_pretrainedmodel_public_dataset/food101/st_efficientnetlcv1_128_tfs/st_efficientnetlcv1_128_tfs_qdq_int8.onnx",
		"description": "ST Efficientnet low complexity model trained from scratch on food101 dataset with 101 classes, 1000 images per class (input 128x128 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "food101",
		"readme": "./image_classification/efficientnet/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/efficientnetv2/ST_pretrainedmodel_public_dataset/food101/efficientnetv2b0_224_fft/efficientnetv2b0_224_fft_qdq_int8.onnx",
		"description": "Efficientnet v2 model trained on food101 dataset with 101 classes, 1000 images per class (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "food101",
		"readme": "./image_classification/efficientnetv2/README.md",
		"visible": true
	},	
	{
		"name": "./image_classification/efficientnetv2/Public_pretrainedmodel_public_dataset/imagenet/efficientnetv2s_384/efficientnetv2s_384_qdq_int8.onnx",
		"description": "Efficientnet v2 pre-trained on imagenet dataset with 1000 classes (input 384x384 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/efficientnetv2/README.md",
		"visible": false
	},	
	{
		"name": "./image_classification/fdmobilenet_pt/Public_pretrainedmodel_public_dataset/Imagenet/fdmobilenet_a025_pt_224/fdmobilenet_a025_pt_224_qdq_int8.onnx",
		"description": "FdMobileNet 0.25 pre-trained on imagenet dataset with 1000 classes",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/fdmobilenet_pt/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/mobilenet_pt/Public_pretrainedmodel_public_dataset/Imagenet/mobilenet_a050_pt_224/mobilenet_a050_pt_224_qdq_int8.onnx",
		"description": "MobileNet 0.5 pre-trained model on imagenet dataset with 1000 classes (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/mobilenet_pt/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/squeezenet_pt/Public_pretrainedmodel_public_dataset/Imagenet/squeezenetv10_pt_224/squeezenetv10_pt_224_qdq_int8.onnx",
		"description": "SqueezeNet v1.0 model trained on imagenet dataset with 1000 classes, (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/squeezenet_pt/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/squeezenetv11/ST_pretrainedmodel_public_dataset/food101/squeezenetv11_224_tfs/squeezenetv11_224_tfs_int8.tflite",
		"description": "SqueezeNet v1.1 trained from scratch on food101 dataset with 101 classes, 1000 images per class (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "food101",
		"readme": "./image_classification/squeezenetv11/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/mobilenetv2/ST_pretrainedmodel_public_dataset/food101/mobilenetv2_a100_224_fft/mobilenetv2_a100_224_fft_qdq_w4_30.91%_w8_69.09%_a8_100%_acc_80.06.onnx",
		"description": "ST MobileNetV2 100 trained from scratch on food101 dataset with 101 classes, 1000 images per class, quantized with mixed precision (w4 30.91%, w8 69.09%, a8 100%) with accuracy 80.06% (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "food101",
		"readme": "./image_classification/mobilenetv2/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/preresnet18_pt/Public_pretrainedmodel_public_dataset/Imagenet/preresnet18_a025_pt_224/preresnet18_a025_pt_224_qdq_int8.onnx",
		"description": "PreResNet18 model pre-trained on imagenet dataset with 1000 classes (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/preresnet18_pt/README.md",
		"visible": true
	},
	{
		"name": "./image_classification/shufflenetv2_pt/Public_pretrainedmodel_public_dataset/Imagenet/shufflenetv2b_x100_pt_224/shufflenetv2b_x100_pt_224_qdq_int8.onnx",
		"description": "ShuffleNetV2 b_x1.0 model pre-trained on imagenet dataset with 1000 classes (input 224x224 rgb image)",
		"usecase": "image_classification",
		"dataset_used": "imagenet",
		"readme": "./image_classification/shufflenetv2_pt/README.md",
		"visible": true
	},
	{
		"name": "./neural_style_transfer/xinet_picasso_muse/Public_pretrainedmodel_public_dataset/coco_2017_80_classes_picasso/xinet_a75_picasso_muse_160/xinet_a75_picasso_muse_160_nomp.tflite",
		"description": "Xinet Picasso Muse model trained on COCO 2017 (80 Classes) with Picasso style (input 160x160 rgb image)",
		"usecase": "neural_style_transfer",
		"dataset_used": "COCO 2017 80 classes",
		"readme": "./neural_style_transfer/xinet_picasso_muse/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/ssd_mobilenetv1_pt/Public_pretrainedmodel_public_dataset/coco/ssd_mobilenetv1_pt_coco_300/ssd_mobilenetv1_pt_coco_300_qdq_int8.onnx",
		"description": "SSD MobileNetV1 model trained on COCO dataset with 80 classes (input 300x300 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO",
		"readme": "./object_detection/ssd_mobilenetv1_pt/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/yolov2t/ST_pretrainedmodel_public_dataset/coco_2017_person/yolov2t_416/yolov2t_416_int8.tflite",
		"description": "tiny yolo v2 model trained on COCO 2017 (Person Detection) with 1 class (input 416x416 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 Person",
		"readme": "./object_detection/yolov2t/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yoloxn/ST_pretrainedmodel_public_dataset/coco_2017_80_classes/st_yoloxn_d100_w025_480/st_yoloxn_d100_w025_480_qdq_int8.onnx",
		"description": "ST yoloxn (derived from yoloxn) trained on COCO 2017 (80 Classes) (input 480x480 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 80 classes",
		"readme": "./object_detection/st_yoloxn/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yoloxn/ST_pretrainedmodel_public_dataset/coco_2017_person/st_yoloxn_d033_w025_416/st_yoloxn_d033_w025_416_qdq_int8.onnx",
		"description": "ST yoloxn (derived from yoloxn) trained on COCO 2017 (Person Detection) with 1 class (input 416x416 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 Person",
		"readme": "./object_detection/st_yoloxn/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yolodv2milli_pt/ST_pretrainedmodel_public_dataset/coco/st_yolodv2milli_actrelu_pt_coco_192/st_yolodv2milli_actrelu_pt_coco_192_qdq_int8.onnx",
		"description": "ST yolodv2milli model trained on COCO 2017 (80 Classes) (input 192x192 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 80 classes",
		"readme": "./object_detection/st_yolodv2milli_pt/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yololcv1/ST_pretrainedmodel_public_dataset/coco_2017_person/st_yololcv1_224/st_yololcv1_224_qdq_int8.onnx",
		"description": "ST yololcv1 model trained on COCO 2017 (Person Detection) with 1 class (input 224x224 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 Person",
		"readme": "./object_detection/st_yololcv1/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yololcv1/ST_pretrainedmodel_public_dataset/coco_2017_person/st_yololcv1_224/st_yololcv1_224_qdq_w4_50.53%_w8_49.47%_a8_100%_map_34.99.onnx",
		"description": "ST yololcv1 model trained on COCO 2017 (Person Detection) with 1 class (input 224x224 rgb image). Quantized with mixed precision (w4 50.53%, w8 49.47%, a8 100%) with mAP 34.99%",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 Person",
		"readme": "./object_detection/st_yololcv1/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/ssdlite_mobilenetv3small_pt/Public_pretrainedmodel_public_dataset/coco/ssdlite_mobilenetv3small_pt_coco_300/ssdlite_mobilenetv3small_pt_coco_300_qdq_int8.onnx",
		"description": "SSDLite MobileNetV3Small model trained on COCO dataset with 80 classes (input 300x300 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 80 classes",
		"readme": "./object_detection/ssdlite_mobilenetv3small_pt/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/ssdlite_mobilenetv3small_pt/Public_pretrainedmodel_public_dataset/voc/ssdlite_mobilenetv3small_pt_voc_300/ssdlite_mobilenetv3small_pt_voc_300_qdq_int8.onnx",
		"description": "SSDLite MobileNetV3Small model trained on VOC dataset with 20 classes (input 300x300 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "VOC 2012 20 classes",
		"readme": "./object_detection/ssdlite_mobilenetv3small_pt/README.md",
		"visible": true
	},
	{
		"name": "./object_detection/st_yoloxn/ST_pretrainedmodel_public_dataset/coco_2017_person/st_yoloxn_d033_w025_416/st_yoloxn_d033_w025_416_int8.tflite",
		"description": "ST yolo X (derived from yoloxn) model trained on COCO 2017 (Person Detection) with 1 class (input 416x416 rgb image)",
		"usecase": "object_detection",
		"dataset_used": "COCO 2017 Person",
		"readme": "./object_detection/st_yoloxn/README.md",
		"visible": true
	},
	{
		"name": "./pose_estimation/movenet/ST_pretrainedmodel_custom_dataset/custom_coco_person_17kpts/st_movenet_lightning_a100_heatmaps_256/st_movenet_lightning_a100_heatmaps_256_int8.tflite",
		"description": "Movenet variant trained on COCO 2017 (Person) with 1 class with 17 key points (input 256x256 rgb image)",
		"usecase": "pose_estimation",
		"dataset_used": "Custom COCO 2017 Person",
		"readme": "./pose_estimation/movenet/README.md",
		"visible": true
	},
	{
		"name": "./re_identification/mobilenetv2/ST_pretrainedmodel_public_dataset/DeepSportradar/mobilenetv2_a100_256_128_fft/mobilenetv2_a100_256_128_fft_int8.tflite",
		"description": "MobileNetV2 variant trained on DeepSportradar dataset (input 256x128 rgb image)",
		"usecase": "re_identification",
		"dataset_used": "DeepSportradar",
		"readme": "./re_identification/mobilenetv2/README.md",
		"visible": true
	},
	{
		"name": "./re_identification/osnet/ST_pretrainedmodel_public_dataset/DeepSportradar/osnet_a025_256_128_tfs/osnet_a025_256_128_tfs_int8.tflite",
		"description": "OSNet variant trained on DeepSportradar dataset (input 256x128 rgb image)",
		"usecase": "re_identification",
		"dataset_used": "DeepSportradar",
		"readme": "./re_identification/osnet/README.md",
		"visible": true
	},
	{
		"name": "./semantic_segmentation/deeplabv3/ST_pretrainedmodel_public_dataset/person_coco_2017_pascal_voc_2012/st_deeplabv3_mnv2_a050_s16_asppv2_256/st_deeplabv3_mnv2_a050_s16_asppv2_256_qdq_int8.onnx",
		"description": "Deeplabv3 variant trained on fused coco 2017 and pascal voc 2012 dataset with person (input 256x256 rgb image)",
		"usecase": "semantic_segmentation",
		"dataset_used": "person_coco_2017_pascal_voc_2012",
		"readme": "./semantic_segmentation/deeplabv3/README.md",
		"visible": true
	},	
	{
		"name": "./semantic_segmentation/deeplabv3/ST_pretrainedmodel_public_dataset/person_coco_2017_pascal_voc_2012/st_deeplabv3_mnv2_a050_s16_asppv2_320/st_deeplabv3_mnv2_a050_s16_asppv2_320_qdq_int8.onnx",
		"description": "Deeplabv3 variant trained on fused coco 2017 and pascal voc 2012 dataset with person (input 320x320 rgb image)",
		"usecase": "semantic_segmentation",
		"dataset_used": "person_coco_2017_pascal_voc_2012",
		"readme": "./semantic_segmentation/deeplabv3/README.md",
		"visible": true
	},
	{
		"name": "./speech_enhancement/stft_tcnn/ST_pretrainedmodel_public_dataset/valentini/stft_tcnn_medium_sigmoid_257x40_qdq_int8.onnx",
		"description": "Speech enhancement to remove audio noise - trained on Valentini dataset",
		"usecase": "speech_enhancement",
		"dataset_used": "Valentini",
		"readme": "./speech_enhancement/stft_tcnn/README.md",
		"visible": true
	}
]