Model eXplorer

At MemryX, we do not maintain a curated model zoo which limits customers choice. Instead, we recommend customers choose/train their own model(s) and use MemryX automated tools to generate executable DFP files (.dfp = “dataflow program”).

Our automated tools do not modify original source models (e.g. no pruning, no distillation, etc.) and do not require any retraining (or any dataset) to maintain high accuracy. Instead, our tools compile models as they were trained, resulting in dataflow programs that execute with high accuracy and high utilization of available hardware.

The list of models below was created using a model explorer that automatically aggregates, compiles, and tests models from various public online sources on MemryX hardware. Anyone can use these publicly available models, access links for detailed model source information, and either recreate or simply download the compiled DFP. None of these models were created by MemryX or belong to MemryX. This model list will evolve over time depending on customer feedback.

Of course, a customer can certainly choose to modify any given model(s) prior to using MemryX tools. For example, if a customer is willing to tradeoff accuracy for performance, one can choose to prune, quantize, and/or distill model(s) and then create an executable for measuring performance & accuracy.

Measurement Information

• SDK Version: 1.1.0

• The host used for these measurements is: Intel i7 9700T (1.5GHz), 16 GB RAM

• The FPS and latency metrics reported here reflect the performance of models running at both Default (14 TFLOPS) and Full (20 TFLOPS) performance modes. Future updates will include benchmarks for configurations up to 40 TFLOPS, providing even more detailed insights into the system’s capabilities.

• Latency ¹: Frame-to-frame latency (used to compute fps)

• Latency ²: Full system latency (including host) from initial input to final output

Models with Tutorials and Examples

Models featured in tutorials carry the badge, indicating code-specific learning paths, while models used in complete example applications are marked with the / badges. Explore in-depth tutorials at Tutorials and discover end-to-end examples at MemryX eXamples.

Performance Metrics: FPS (14 TFLOPS) FPS (20 TFLOPS) Latency (14 TFLOPS) Latency (20 TFLOPS)

Models (Below is a partial list of tested models)

Below is a partial list of AI models that MemryX has downloaded, compiled, and benchmarked from public sources. This is not a model zoo; the models remain unmodified, and MemryX customers are not restricted to using these models. Customers are welcome to compile additional models as needed. For guidance, see the Neural Compiler and Tutorials pages.

Classification (ImageNet)

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
1	CSP-DarkNet50	256×256×3	26.3 M	ONNX	468	662	2.1 ms	1.5 ms	9.9 ms	7.2 ms	OpenMMLab	DFP
2	CSP-ResNet-50	256×256×3	20.6 M	ONNX	869	1127	1.2 ms	0.9 ms	5.4 ms	4.1 ms	OpenMMLab	DFP
3	DLA34	224×224×3	15.0 M	ONNX	603	854	1.7 ms	1.2 ms	4.4 ms	3.2 ms	OpenVINO	DFP
4	DenseNet121	224×224×3	7.7 M	Keras	711	1008	1.4 ms	1.0 ms	11.1 ms	8.0 ms	Keras.Applications	DFP
5	DenseNet169	224×224×3	13.6 M	Keras	521	738	1.9 ms	1.4 ms	18.9 ms	13.5 ms	Keras.Applications	DFP
6	DenseNet201	224×224×3	19.3 M	Keras	334	473	3.0 ms	2.1 ms	32.1 ms	22.8 ms	Keras.Applications	DFP
7	EfficientNet-B0	224×224×3	5.0 M	ONNX	761	1078	1.3 ms	0.9 ms	8.0 ms	5.9 ms	OpenVINO	DFP
8	EfficientNet-B1	240×240×3	7.4 M	ONNX	455	644	2.2 ms	1.6 ms	16.6 ms	12.1 ms	OpenMMLab	DFP
9	EfficientNet-EdgeTpu(L)	300×300×3	10.1 M	TFLite	166	235	6.0 ms	4.3 ms	14.5 ms	10.4 ms	Coral	DFP
10	EfficientNet-EdgeTpu(M)	240×240×3	5.3 M	TFLite	450	637	2.2 ms	1.6 ms	6.5 ms	4.8 ms	Coral	DFP
11	EfficientNet-EdgeTpu(S)	224×224×3	5.2 M	TFLite	488	691	2.0 ms	1.4 ms	4.3 ms	3.2 ms	Coral	DFP
12	EfficientNet-lite0	224×224×3	4.4 M	TFLite	1935	2304	0.5 ms	0.4 ms	2.1 ms	1.6 ms	TFHub	DFP
13	EfficientNet-lite1	240×240×3	5.1 M	TFLite	1929	2295	0.5 ms	0.4 ms	2.5 ms	2.0 ms	TFHub	DFP
14	EfficientNet-lite2	260×260×3	5.8 M	TFLite	1537	1780	0.7 ms	0.6 ms	3.4 ms	2.7 ms	TFHub	DFP
15	EfficientNet-lite3	280×280×3	7.8 M	TFLite	1218	1724	0.8 ms	0.6 ms	4.3 ms	3.2 ms	TFHub	DFP
16	EfficientNet-lite4	300×300×3	12.4 M	TFLite	750	1062	1.3 ms	0.9 ms	8.9 ms	6.5 ms	TFHub	DFP
17	EfficientNet-v2-B0	224×224×3	6.8 M	ONNX	1101	1558	0.9 ms	0.6 ms	6.5 ms	4.8 ms	OpenVINO	DFP
18	EfficientNet-v2-B0	224×224×3	6.9 M	Keras	1102	1560	0.9 ms	0.6 ms	6.0 ms	4.4 ms	Keras.Applications	DFP
19	EfficientNet-v2-B1	240×240×3	7.7 M	ONNX	1015	1437	1.0 ms	0.7 ms	10.3 ms	7.5 ms	OpenMMLab	DFP
20	EfficientNet-v2-B1	240×240×3	7.8 M	Keras	1019	1442	1.0 ms	0.7 ms	8.7 ms	6.3 ms	Keras.Applications	DFP
21	EfficientNet-v2-B2	260×260×3	9.6 M	ONNX	761	1077	1.3 ms	0.9 ms	14.7 ms	10.5 ms	OpenMMLab	DFP
22	EfficientNet-v2-B2	260×260×3	9.7 M	Keras	763	1080	1.3 ms	0.9 ms	13.4 ms	9.6 ms	Keras.Applications	DFP
23	EfficientNet-v2-B3	300×300×3	13.8 M	Keras	524	742	1.9 ms	1.3 ms	25.4 ms	18.1 ms	Keras.Applications	DFP
24	GoogleNet	224×224×3	6.7 M	ONNX	519	734	1.9 ms	1.4 ms	3.9 ms	2.9 ms	ONNX Zoo	DFP
25	HBONet0.25	224×224×3	1.8 M	ONNX	1938	2303	0.5 ms	0.4 ms	1.4 ms	1.2 ms	OpenVINO	DFP
26	HBONet1.0	224×224×3	4.3 M	ONNX	1113	1178	0.9 ms	0.8 ms	3.2 ms	2.6 ms	OpenVINO	DFP
27	Inception	224×224×3	6.3 M	TFLite	753	1066	1.3 ms	0.9 ms	3.0 ms	2.3 ms	Coral	DFP
28	Inception	224×224×3	6.7 M	ONNX	537	760	1.9 ms	1.3 ms	3.7 ms	2.8 ms	ONNX Zoo	DFP
29	Inception-ResNet-v2	299×299×3	28.9 M	TensorFlow	168	238	6.0 ms	4.2 ms	36.5 ms	26.0 ms	OpenVINO	DFP
30	Inception-v2	224×224×3	10.7 M	ONNX	751	1064	1.3 ms	0.9 ms	3.9 ms	2.9 ms	ONNX Zoo	DFP
31	Inception-v3	299×299×3	22.7 M	ONNX	318	450	3.1 ms	2.2 ms	10.1 ms	7.2 ms	OpenVINO	DFP
32	Inception-v3	299×299×3	22.7 M	TFLite	318	450	3.1 ms	2.2 ms	10.1 ms	7.3 ms	Coral	DFP
33	Inception-v3	346×346×3	22.7 M	TFLite	276	391	3.6 ms	2.6 ms	12.0 ms	8.7 ms	AI Benchmark	DFP
34	Inception-v3	299×299×3	22.7 M	TensorFlow	318	450	3.1 ms	2.2 ms	10.1 ms	7.3 ms	OpenVINO	DFP
35	Inception-v3	224×224×3	10.7 M	TFLite	750	1063	1.3 ms	0.9 ms	3.7 ms	2.8 ms	Coral	DFP
36	Inception-v3	299×299×3	22.7 M	Keras	318	450	3.1 ms	2.2 ms	10.1 ms	7.2 ms	Keras.Applications	DFP
37	MobileNet	224×224×3	4.0 M	ONNX	1946	2303	0.5 ms	0.4 ms	1.5 ms	1.2 ms	OpenCV	DFP
38	MobileNet	128×128×3	4.0 M	TFLite	6771	7839	0.1 ms	0.1 ms	1.1 ms	0.8 ms	TFHub	DFP
39	MobileNet	192×192×3	4.0 M	TFLite	3579	4210	0.3 ms	0.2 ms	1.4 ms	1.1 ms	TFHub	DFP
40	MobileNet	224×224×3	4.0 M	TFLite	1932	2301	0.5 ms	0.4 ms	1.7 ms	1.4 ms	TFHub	DFP
41	MobileNet	160×160×3	4.0 M	TFLite	4248	4836	0.2 ms	0.2 ms	1.3 ms	1.0 ms	TFHub	DFP
42	MobileNet	224×224×3	4.1 M	TensorFlow	1933	2301	0.5 ms	0.4 ms	1.6 ms	1.3 ms	OpenVINO	DFP
43	MobileNet	224×224×3	4.1 M	Keras	1935	2307	0.5 ms	0.4 ms	1.5 ms	1.2 ms	Keras.Applications	DFP
44	MobileNet-EdgeTPU-v2-extra-small	224×224×3	4.2 M	TFLite	1938	2271	0.5 ms	0.4 ms	2.1 ms	1.7 ms	TFHub	DFP
45	MobileNet-v2	224×224×3	3.3 M	ONNX	1938	2302	0.5 ms	0.4 ms	1.5 ms	1.2 ms	OpenVINO	DFP
46	MobileNet-v2	224×224×3	3.3 M	TFLite	1933	2311	0.5 ms	0.4 ms	1.6 ms	1.3 ms	TFHub	DFP
47	MobileNet-v2	224×224×3	3.4 M	Keras	1935	2302	0.5 ms	0.4 ms	1.5 ms	1.2 ms	Keras.Applications	DFP
48	MobileNet-v3	224×224×3	3.9 M	TFLite	1937	2297	0.5 ms	0.4 ms	1.9 ms	1.5 ms	Coral	DFP
49	MobileNet-v3	512×512×3	5.2 M	TFLite	454	577	2.2 ms	1.7 ms	9.6 ms	7.2 ms	AI Benchmark	DFP
50	MobileOne	224×224×3	7.4 M	ONNX	1927	2295	0.5 ms	0.4 ms	2.3 ms	1.8 ms	GitHub/PINTO	DFP
51	NasNet(Mobile)	224×224×3	4.2 M	TFLite	1931	2312	0.5 ms	0.4 ms	1.9 ms	1.6 ms	TFHub	DFP
52	NasNet(Mobile)	160×160×3	4.2 M	TFLite	4211	4831	0.2 ms	0.2 ms	1.5 ms	1.2 ms	TFHub	DFP
53	NasNet(Mobile)	192×192×3	4.2 M	TFLite	3570	4208	0.3 ms	0.2 ms	1.7 ms	1.4 ms	TFHub	DFP
54	NasNet(Mobile)	128×128×3	4.2 M	TFLite	4731	6692	0.2 ms	0.1 ms	1.4 ms	1.1 ms	TFHub	DFP
55	NasNet(Mobile)-1.3	224×224×3	5.9 M	TFLite	1930	2296	0.5 ms	0.4 ms	2.3 ms	1.8 ms	TFHub	DFP
56	PP-LCNet-v2	224×224×3	6.3 M	TensorFlow	1934	2294	0.5 ms	0.4 ms	2.7 ms	2.1 ms	GitHub/PINTO	DFP
57	PpResNet-50	224×224×3	24.4 M	ONNX	742	1050	1.3 ms	1.0 ms	6.0 ms	4.5 ms	OpenCV	DFP
58	RegNetX002	224×224×3	2.6 M	Keras	1934	2304	0.5 ms	0.4 ms	2.0 ms	1.6 ms	Keras.Applications	DFP
59	RegNetX004	224×224×3	5.0 M	Keras	1858	2266	0.5 ms	0.4 ms	3.4 ms	2.5 ms	Keras.Applications	DFP
60	RegNetX006	224×224×3	5.9 M	Keras	1199	1697	0.8 ms	0.6 ms	3.9 ms	2.9 ms	Keras.Applications	DFP
61	RegNetX008	224×224×3	7.0 M	Keras	1003	1420	1.0 ms	0.7 ms	4.7 ms	3.5 ms	Keras.Applications	DFP
62	RegNetX016	224×224×3	8.8 M	Keras	290	410	3.4 ms	2.4 ms	12.3 ms	8.8 ms	Keras.Applications	DFP
63	RegNetY002	224×224×3	3.0 M	Keras	1921	2289	0.5 ms	0.4 ms	3.8 ms	2.9 ms	Keras.Applications	DFP
64	RegNetY004	224×224×3	4.2 M	Keras	773	1094	1.3 ms	0.9 ms	10.2 ms	7.4 ms	Keras.Applications	DFP
65	RegNetY006	224×224×3	5.8 M	Keras	1118	1583	0.9 ms	0.6 ms	9.4 ms	6.8 ms	Keras.Applications	DFP
66	RegNetY008	224×224×3	6.0 M	Keras	1286	1763	0.8 ms	0.6 ms	7.9 ms	5.8 ms	Keras.Applications	DFP
67	RegNetY016	224×224×3	10.7 M	Keras	220	312	4.5 ms	3.2 ms	57.3 ms	40.6 ms	Keras.Applications	DFP
68	RepVGG-A0	224×224×3	7.9 M	ONNX	1839	2298	0.5 ms	0.4 ms	2.1 ms	1.6 ms	OpenMMLab	DFP
69	RepVGG-A1	224×224×3	12.2 M	ONNX	1928	2296	0.5 ms	0.4 ms	2.2 ms	1.7 ms	OpenMMLab	DFP
70	RepVGG-A2	224×224×3	24.3 M	ONNX	742	1051	1.3 ms	1.0 ms	4.9 ms	3.5 ms	OpenMMLab	DFP
71	RepVGG-B0	224×224×3	13.7 M	ONNX	1079	1527	0.9 ms	0.7 ms	3.2 ms	2.4 ms	OpenMMLab	DFP
72	ResNet-18	224×224×3	11.1 M	ONNX	1824	2237	0.5 ms	0.4 ms	1.8 ms	1.4 ms	OpenVINO	DFP
73	ResNet-18-v2	224×224×3	11.2 M	ONNX	1823	2223	0.5 ms	0.4 ms	1.9 ms	1.5 ms	ONNX Zoo	DFP
74	ResNet-34	224×224×3	20.8 M	ONNX	888	1258	1.1 ms	0.8 ms	4.5 ms	3.4 ms	OpenVINO	DFP
75	ResNet-34-v2	224×224×3	20.8 M	ONNX	967	1369	1.0 ms	0.7 ms	4.5 ms	3.3 ms	ONNX Zoo	DFP
76	ResNet-50	224×224×3	24.4 M	ONNX	826	1169	1.2 ms	0.9 ms	6.2 ms	4.6 ms	ONNX Zoo	DFP
77	ResNet-50	224×224×3	24.3 M	TFLite	811	1148	1.2 ms	0.9 ms	4.9 ms	3.6 ms	Coral	DFP
78	ResNet-50	224×224×3	24.4 M	Keras	811	1148	1.2 ms	0.9 ms	5.2 ms	3.8 ms	Keras.Applications	DFP
79	ResNet-50-v2	224×224×3	24.4 M	Keras	820	1162	1.2 ms	0.9 ms	5.7 ms	4.2 ms	Keras.Applications	DFP
80	ShuffleNet	224×224×3	1.4 M	ONNX	1943	2297	0.5 ms	0.4 ms	2.0 ms	1.6 ms	ONNX Zoo	DFP
81	ShuffleNet-v2	224×224×3	2.2 M	ONNX	1937	2301	0.5 ms	0.4 ms	1.7 ms	1.4 ms	ONNX Zoo	DFP
82	SqueezeNet	224×224×3	1.2 M	TFLite	1036	1466	1.0 ms	0.7 ms	1.8 ms	1.4 ms	TFHub	DFP
83	SqueezeNet	224×224×3	1.2 M	ONNX	1953	2316	0.5 ms	0.4 ms	1.2 ms	1.1 ms	ONNX Zoo	DFP
84	SqueezeNet1.1	224×224×3	1.2 M	ONNX	1937	2322	0.5 ms	0.4 ms	1.3 ms	1.1 ms	ONNX Zoo	DFP
85	Xception	299×299×3	21.8 M	Keras	300	425	3.3 ms	2.4 ms	13.0 ms	9.3 ms	Keras.Applications	DFP
86	YOLO-v8-large-CLS	224×224×3	35.7 M	TFLite	259	366	3.9 ms	2.7 ms	13.4 ms	9.6 ms	Ultralytics	DFP
87	YOLO-v8-medium-CLS	224×224×3	16.3 M	TFLite	1018	1441	1.0 ms	0.7 ms	4.4 ms	3.2 ms	Ultralytics	DFP
88	YOLO-v8-nano-CLS	224×224×3	2.6 M	TFLite	1937	2316	0.5 ms	0.4 ms	1.4 ms	1.1 ms	Ultralytics	DFP
89	YOLO-v8-small-CLS	224×224×3	6.1 M	TFLite	1931	2297	0.5 ms	0.4 ms	2.0 ms	1.6 ms	Ultralytics	DFP

Object Detection

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
90	CenterNet(DLA)	512×512×3	17.1 M	ONNX	107	151	9.3 ms	6.6 ms	23.8 ms	17.0 ms	OpenVINO	DFP
91	CenterNet-MobileNet-v2	320×320×3	2.3 M	TensorFlow	697	950	1.4 ms	1.1 ms	4.4 ms	3.6 ms	TF Object Detection Zoo	DFP
92	CenterNet-MobileNet-v2-kpts	320×320×3	2.3 M	ONNX	462	476	2.2 ms	2.1 ms	5.2 ms	4.6 ms	TF Object Detection Zoo	DFP
93	CenterNet-ResNet-fpn	512×512×3	24.2 M	TensorFlow	67	94	14.9 ms	10.6 ms	27.8 ms	20.4 ms	TF Object Detection Zoo	DFP
94	DAMO-YOLO-Nl	416×416×3	5.4 M	ONNX	312	443	3.2 ms	2.3 ms	10.0 ms	7.3 ms	DaMoYolo	DFP
95	DAMO-YOLO-Ns	416×416×3	1.3 M	ONNX	611	865	1.6 ms	1.2 ms	4.7 ms	3.6 ms	DaMoYolo	DFP
96	DAMO-YOLO-small	640×640×3	15.5 M	ONNX	144	204	6.9 ms	4.9 ms	18.9 ms	13.7 ms	DaMoYolo	DFP
97	DAMO-YOLO-tiny	640×640×3	8.1 M	ONNX	386	386	2.6 ms	2.6 ms	10.7 ms	10.7 ms	DaMoYolo	DFP
98	DbNet-ResNet-18	736×1344×3	11.6 M	ONNX	17	24	58.8 ms	41.7 ms	75.5 ms	53.6 ms	OpenMMLab	DFP
99	DeepSort	128×64×3	10.6 M	TensorFlow	2016	2854	0.5 ms	0.4 ms	1.7 ms	1.3 ms	GitHub/PINTO	DFP
100	DroNet	608×608×3	67.4 K	TFLite	364	445	2.7 ms	2.2 ms	4.0 ms	3.2 ms	GitHub/PINTO	DFP
101	EfficientDet-Lite1	384×384×3	4.2 M	TFLite	223	230	4.5 ms	4.3 ms	16.8 ms	14.2 ms	TFHub	DFP
102	MediaPipe-object-Detection-3d-camera	224×224×3	2.4 M	TFLite	1922	2298	0.5 ms	0.4 ms	2.0 ms	1.6 ms	MediaPipe	DFP
103	MediaPipe-object-Detection-3d-chair-1stag…	640×480×3	1.4 M	TFLite	412	502	2.4 ms	2.0 ms	5.4 ms	4.3 ms	MediaPipe	DFP
104	MediaPipe-object-Detection-3d-cup	224×224×3	2.4 M	TFLite	1927	2291	0.5 ms	0.4 ms	2.0 ms	1.6 ms	MediaPipe	DFP
105	MediaPipe-object-Detection-3d-sneakers	224×224×3	2.4 M	TFLite	1929	2306	0.5 ms	0.4 ms	2.0 ms	1.6 ms	MediaPipe	DFP
106	MediaPipe-object-Detection-3d-sneakers-1s…	640×480×3	1.4 M	TFLite	411	502	2.4 ms	2.0 ms	5.5 ms	4.3 ms	MediaPipe	DFP
107	Mobile-object-labeler	224×224×3	3.5 M	TFLite	1936	2306	0.5 ms	0.4 ms	1.5 ms	1.2 ms	TFHub	DFP
108	Mobile-object-localizer	192×192×3	1.7 M	TFLite	3577	4214	0.3 ms	0.2 ms	1.3 ms	1.1 ms	GitHub/PINTO	DFP
109	MobileNet-YOLO-v4-syg	416×416×3	10.0 M	TensorFlow	766	956	1.3 ms	1.0 ms	8.1 ms	6.0 ms	OpenVINO	DFP
110	NanoDet	416×416×3	919.8 K	ONNX	785	947	1.3 ms	1.1 ms	3.9 ms	3.3 ms	OpenCV	DFP
111	NanoDet+ (ShuffleNet)-v2-1.5x	416×416×3	2.3 M	ONNX	617	767	1.6 ms	1.3 ms	6.2 ms	5.0 ms	OpenVINO	DFP
112	NanoDetwithShuffleNet-v2-1.5x	416×416×3	2.0 M	ONNX	694	891	1.4 ms	1.1 ms	4.8 ms	3.8 ms	OpenVINO	DFP
113	Person-Detection	512×512×3	1.7 M	TFLite	475	580	2.1 ms	1.7 ms	3.9 ms	3.2 ms	GitHub/PINTO	DFP
114	Person-vehicle-bike-Detection-crossroad-0…	1024×1024×3	1.1 M	TFLite	117	156	8.5 ms	6.4 ms	14.1 ms	10.9 ms	GitHub/PINTO	DFP
115	Person-vehicle-bike-Detection-crossroad-1…	512×512×3	2.8 M	TFLite	475	578	2.1 ms	1.7 ms	4.4 ms	3.5 ms	GitHub/PINTO	DFP
116	SSD-FPNMobileNet	640×640×3	29.5 M	TFLite	41	58	24.4 ms	17.2 ms	44.7 ms	34.0 ms	Coral	DFP
117	SSD-MobileNet	300×300×3	6.5 M	TFLite	1196	1436	0.8 ms	0.7 ms	2.1 ms	1.8 ms	Coral	DFP
118	SSD-MobileNet-v2	300×300×3	5.8 M	TFLite	1197	1423	0.8 ms	0.7 ms	2.5 ms	2.0 ms	Coral	DFP
119	SSD-MobileNet-v2 (MNAS-FPN)	320×320×3	2.4 M	TFLite	307	323	3.3 ms	3.1 ms	8.5 ms	7.5 ms	GitHub/PINTO	DFP
120	SSD512	512×512×3	34.4 M	ONNX	21	29	47.6 ms	34.5 ms	65.7 ms	46.6 ms	OpenMMLab	DFP
121	SSDLiteMobileDet	320×320×3	3.9 M	TFLite	1189	1420	0.8 ms	0.7 ms	3.1 ms	2.4 ms	Coral	DFP
122	SSDlite-MobileNet-v2	320×320×3	4.0 M	ONNX	1192	1423	0.8 ms	0.7 ms	2.6 ms	2.1 ms	OpenMMLab	DFP
123	SSDlite-MobileNet-v2	300×300×3	4.3 M	TensorFlow	1191	1426	0.8 ms	0.7 ms	2.5 ms	1.9 ms	GitHub/PINTO	DFP
124	SpaghettiNet-EdgeTPU	320×320×3	3.3 M	TFLite	1182	1419	0.8 ms	0.7 ms	3.4 ms	2.8 ms	GitHub/PINTO	DFP
125	Text-Detection-db	480×640×3	11.6 M	TensorFlow	59	84	16.9 ms	11.9 ms	22.8 ms	17.1 ms	GitHub/PINTO	DFP
126	Text-recognition-CRNN-EN8	32×100×1	5.3 M	ONNX	3304	4675	0.3 ms	0.2 ms	1.4 ms	1.1 ms	OpenCV	DFP
127	Vehicle-license-plate-Detection-barrier-0…	300×300×3	619.4 K	TFLite	1202	1449	0.8 ms	0.7 ms	1.6 ms	1.3 ms	GitHub/PINTO	DFP
128	VehicleLicensePlateDetectionBarrier	256×256×3	561.2 K	TensorFlow	1675	2008	0.6 ms	0.5 ms	2.4 ms	1.9 ms	OpenVINO	DFP
129	YOLO-tiny	416×416×3	15.1 M	TensorFlow	492	697	2.0 ms	1.4 ms	4.0 ms	3.1 ms	GitHub/RedisAI	DFP
130	YOLO-v2-tiny	416×416×3	15.1 M	TFLite	594	840	1.7 ms	1.2 ms	3.4 ms	2.6 ms	OpenVINO	DFP
131	YOLO-v3-MobileNet-v2	320×320×3	3.2 M	ONNX	713	1009	1.4 ms	1.0 ms	4.9 ms	3.9 ms	OpenMMLab	DFP
132	YOLO-v3-MobileNet-v2-ms	416×416×3	3.2 M	ONNX	546	612	1.8 ms	1.6 ms	6.6 ms	5.3 ms	OpenMMLab	DFP
133	YOLO-v3-lite	256×256×3	6.4 M	ONNX	1665	2006	0.6 ms	0.5 ms	2.8 ms	2.2 ms	GitHub/PINTO	DFP
134	YOLO-v4-tiny	416×416×3	5.8 M	TFLite	608	675	1.6 ms	1.5 ms	3.7 ms	3.1 ms	AI Benchmark	DFP
135	YOLO-v5-medium-leaky	416×416×3	20.2 M	TensorFlow	261	358	3.8 ms	2.8 ms	11.3 ms	8.2 ms	Ultralytics	DFP
136	YOLO-v5-medium-leaky	640×640×3	20.2 M	TensorFlow	120	170	8.3 ms	5.9 ms	21.3 ms	15.9 ms	Ultralytics	DFP
137	YOLO-v5-medium-v6.1-fast	640×640×3	20.2 M	ONNX	118	168	8.5 ms	6.0 ms	23.0 ms	16.6 ms	OpenMMLab	DFP
138	YOLO-v5-medium-v6.1-fastb64-VOC	512×512×3	20.0 M	ONNX	187	256	5.3 ms	3.9 ms	15.6 ms	11.2 ms	OpenMMLab	DFP
139	YOLO-v5-mu	640×640×3	23.9 M	ONNX	93	132	10.8 ms	7.6 ms	28.3 ms	20.2 ms	Ultralytics	DFP
140	YOLO-v5-n6u	1280×1280×3	4.1 M	ONNX	83	98	12.0 ms	10.2 ms	34.4 ms	29.9 ms	Ultralytics	DFP
141	YOLO-v5-nano-leaky	416×416×3	1.8 M	TensorFlow	685	685	1.5 ms	1.5 ms	4.7 ms	4.7 ms	Ultralytics	DFP
142	YOLO-v5-nano-leaky	640×640×3	1.8 M	TensorFlow	273	278	3.7 ms	3.6 ms	9.1 ms	8.5 ms	Ultralytics	DFP
143	YOLO-v5-nano-leaky	320×320×3	1.8 M	TensorFlow	1184	1268	0.8 ms	0.8 ms	3.0 ms	2.7 ms	Ultralytics	DFP
144	YOLO-v5-nano-p6-v6.2-fast	1280×1280×3	3.1 M	ONNX	65	68	15.4 ms	14.7 ms	44.1 ms	40.9 ms	OpenMMLab	DFP
145	YOLO-v5-nano-v6.1-fast	640×640×3	1.8 M	ONNX	266	278	3.8 ms	3.6 ms	10.4 ms	8.7 ms	OpenMMLab	DFP
146	YOLO-v5-nano-v6.1-fastb64-VOC	512×512×3	1.7 M	ONNX	475	573	2.1 ms	1.7 ms	4.3 ms	3.7 ms	OpenMMLab	DFP
147	YOLO-v5-nu	640×640×3	2.5 M	ONNX	326	388	3.1 ms	2.6 ms	7.9 ms	6.6 ms	Ultralytics	DFP
148	YOLO-v5-small-leaky	320×320×3	6.9 M	TensorFlow	1086	1237	0.9 ms	0.8 ms	3.6 ms	3.1 ms	Ultralytics	DFP
149	YOLO-v5-small-leaky	416×416×3	6.9 M	TensorFlow	646	701	1.5 ms	1.4 ms	5.4 ms	4.7 ms	Ultralytics	DFP
150	YOLO-v5-small-leaky	640×640×3	6.9 M	TensorFlow	274	274	3.6 ms	3.6 ms	10.6 ms	10.6 ms	Ultralytics	DFP
151	YOLO-v5-small-p6-v6.2-fast	1280×1280×3	12.0 M	ONNX	42	59	23.8 ms	16.9 ms	58.8 ms	48.4 ms	OpenMMLab	DFP
152	YOLO-v5-small-v6.1-fast	640×640×3	6.9 M	ONNX	264	273	3.8 ms	3.7 ms	12.2 ms	10.1 ms	OpenMMLab	DFP
153	YOLO-v5-small-v6.1-fastb64-VOC	512×512×3	6.7 M	ONNX	473	571	2.1 ms	1.8 ms	5.6 ms	4.3 ms	OpenMMLab	DFP
154	YOLO-v5-su	640×640×3	8.7 M	ONNX	266	376	3.8 ms	2.7 ms	10.2 ms	8.1 ms	Ultralytics	DFP
155	YOLO-v6-medium-fast	640×640×3	32.7 M	ONNX	44	63	22.7 ms	15.9 ms	56.9 ms	40.3 ms	OpenMMLab	DFP
156	YOLO-v6-nano-fast	640×640×3	4.1 M	ONNX	327	392	3.1 ms	2.6 ms	7.2 ms	5.9 ms	OpenMMLab	DFP
157	YOLO-v6-small-fast	640×640×3	16.4 M	ONNX	140	198	7.1 ms	5.1 ms	16.7 ms	12.1 ms	OpenMMLab	DFP
158	YOLO-v6-tiny-fast	640×640×3	9.2 M	ONNX	226	320	4.4 ms	3.1 ms	10.8 ms	8.1 ms	OpenMMLab	DFP
159	YOLO-v7	480×640×3	5.9 M	ONNX	387	407	2.6 ms	2.5 ms	9.5 ms	7.8 ms	GitHub/PINTO	DFP
160	YOLO-v7	256×640×3	35.2 M	ONNX	58	82	17.2 ms	12.2 ms	54.8 ms	38.9 ms	GitHub/PINTO	DFP
161	YOLO-v7-tiny	640×640×3	5.9 M	ONNX	275	277	3.6 ms	3.6 ms	10.5 ms	9.8 ms	GitHub/Official YOLOv7	DFP
162	YOLO-v7-tiny	320×320×3	5.9 M	ONNX	1181	1210	0.8 ms	0.8 ms	3.5 ms	3.0 ms	GitHub/Official YOLOv7	DFP
163	YOLO-v7-tiny	416×416×3	5.9 M	ONNX	660	686	1.5 ms	1.5 ms	5.1 ms	4.6 ms	GitHub/Official YOLOv7	DFP
164	YOLO-v7-tiny-fast-b	640×640×3	5.9 M	ONNX	266	271	3.8 ms	3.7 ms	11.1 ms	9.9 ms	OpenMMLab	DFP
165	YOLO-v8-medium	640×640×3	24.7 M	TFLite	58	81	17.2 ms	12.3 ms	47.0 ms	33.9 ms	Ultralytics	DFP
166	YOLO-v8-medium	640×640×3	24.7 M	ONNX	45	63	22.2 ms	15.9 ms	36.3 ms	26.0 ms	Ultralytics	DFP
167	YOLO-v8-nano	640×640×3	3.0 M	TFLite	326	389	3.1 ms	2.6 ms	8.0 ms	7.0 ms	Ultralytics	DFP
168	YOLO-v8-nano	640×640×3	3.0 M	ONNX	327	388	3.1 ms	2.6 ms	8.3 ms	6.9 ms	Ultralytics	DFP
169	YOLO-v8-nano-Oriented-Bounding-Boxes	1024×1024×3	2.9 M	ONNX	128	153	7.8 ms	6.5 ms	16.6 ms	13.6 ms	Ultralytics	DFP
170	YOLO-v8-small	640×640×3	10.6 M	TFLite	229	323	4.4 ms	3.1 ms	12.6 ms	9.8 ms	Ultralytics	DFP
171	YOLO-v8-small	640×640×3	10.6 M	ONNX	223	316	4.5 ms	3.2 ms	11.6 ms	8.6 ms	Ultralytics	DFP
172	YOLO-v8-small-Oriented-Bounding-Boxes	1024×1024×3	10.9 M	ONNX	75	104	13.3 ms	9.6 ms	22.0 ms	16.6 ms	Ultralytics	DFP
173	YOLO-v9-medium	640×640×3	19.1 M	ONNX	38	53	26.3 ms	18.9 ms	58.9 ms	41.6 ms	Ultralytics	DFP
174	YOLO-v9-small	640×640×3	6.9 M	ONNX	110	153	9.1 ms	6.5 ms	32.8 ms	23.3 ms	Ultralytics	DFP
175	YOLO-v9-tiny	640×640×3	2.0 M	ONNX	262	370	3.8 ms	2.7 ms	16.7 ms	12.2 ms	Ultralytics	DFP
176	YOLOX	640×640×3	8.5 M	ONNX	262	372	3.8 ms	2.7 ms	10.1 ms	7.6 ms	OpenCV	DFP
177	YOLOX-medium-ti-lite	416×416×3	24.1 M	ONNX	158	224	6.3 ms	4.5 ms	19.5 ms	14.0 ms	Texas Instruments	DFP
178	YOLOX-medium-ti-lite	640×640×3	24.1 M	ONNX	72	102	13.9 ms	9.8 ms	32.1 ms	23.0 ms	Texas Instruments	DFP
179	YOLOX-nano	416×416×3	883.3 K	ONNX	808	840	1.2 ms	1.2 ms	3.5 ms	3.1 ms	GitHub/YOLOX	DFP
180	YOLOX-nano-ti-lite	416×416×3	2.1 M	ONNX	785	949	1.3 ms	1.1 ms	3.4 ms	3.0 ms	Texas Instruments	DFP
181	YOLOX-small-ti-lite	640×640×3	8.5 M	ONNX	236	321	4.2 ms	3.1 ms	9.8 ms	8.1 ms	Texas Instruments	DFP
182	YOLOX-small-ti-lite	416×416×3	8.5 M	ONNX	510	723	2.0 ms	1.4 ms	5.4 ms	4.2 ms	Texas Instruments	DFP
183	YOLOX-tiny	416×416×3	4.8 M	ONNX	644	837	1.6 ms	1.2 ms	4.9 ms	4.0 ms	OpenVINO	DFP
184	YOLOX-tiny-ti-lite	416×416×3	4.8 M	ONNX	784	951	1.3 ms	1.1 ms	4.5 ms	3.6 ms	Texas Instruments	DFP

Segmentation

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
185	Autoseg-EdgeTPU-small	512×512×3	2.4 M	TFLite	47	66	21.3 ms	15.2 ms	38.2 ms	27.3 ms	TFHub	DFP
186	BiseNet (ResNet-18)	512×512×3	12.7 M	ONNX	126	179	7.9 ms	5.6 ms	17.0 ms	12.6 ms	OpenMMLab	DFP
187	BiseNet (ResNet-18) - CityScapes	512×512×3	12.7 M	ONNX	134	189	7.5 ms	5.3 ms	15.8 ms	11.5 ms	OpenMMLab	DFP
188	BiseNetv2 (FCN) - CityScapes	512×512×3	3.2 M	ONNX	123	175	8.1 ms	5.7 ms	15.0 ms	10.8 ms	OpenMMLab	DFP
189	DeepLab-v3 (MobileNet-v2)	513×513×3	672.2 K	TFLite	475	579	2.1 ms	1.7 ms	4.2 ms	3.4 ms	Coral	DFP
190	Deeplab-EdgeTPU-extra-small	512×512×3	3.1 M	TFLite	411	578	2.4 ms	1.7 ms	5.9 ms	4.6 ms	TFHub	DFP
191	Deeplab-EdgeTPU-medium	512×512×3	6.8 M	TFLite	234	331	4.3 ms	3.0 ms	10.3 ms	7.5 ms	TFHub	DFP
192	Deeplab-EdgeTPU-small	512×512×3	4.5 M	TFLite	322	456	3.1 ms	2.2 ms	7.0 ms	5.2 ms	TFHub	DFP
193	Deeplab-v3	257×257×3	672.2 K	TFLite	1408	1991	0.7 ms	0.5 ms	2.7 ms	2.1 ms	TFHub	DFP
194	Deeplab-v3-plus	1024×1024×3	2.1 M	TFLite	23	29	43.5 ms	34.5 ms	62.2 ms	48.4 ms	AI Benchmark	DFP
195	EdgeTPU-DeepLab-slim	513×513×3	2.0 M	TFLite	474	578	2.1 ms	1.7 ms	5.9 ms	4.8 ms	Coral	DFP
196	Edgetpu-segmentation-EdgeTPU	512×512×3	2.2 M	TFLite	47	67	21.3 ms	14.9 ms	36.8 ms	26.4 ms	GitHub/PINTO	DFP
197	Edgetpu-segmentation-EdgeTPU	16×16×320	49.5 K	TFLite	2436	2643	0.4 ms	0.4 ms	1.2 ms	1.1 ms	GitHub/PINTO	DFP
198	FPN (ResNet-50) - ADE20K	512×512×3	27.2 M	ONNX	19	26	52.6 ms	38.5 ms	72.3 ms	52.2 ms	OpenMMLab	DFP
199	FastSAM-small	640×640×3	11.2 M	ONNX	80	113	12.5 ms	8.8 ms	21.5 ms	15.7 ms	Ultralytics	DFP
200	HRNet	256×192×3	27.2 M	ONNX	108	153	9.3 ms	6.5 ms	46.1 ms	32.7 ms	GitHub/PINTO	DFP
201	Human-segmentation	192×192×3	1.5 M	TFLite	652	746	1.5 ms	1.3 ms	3.3 ms	2.8 ms	GitHub/PINTO	DFP
202	MagicTouch	512×512×4	2.9 M	TFLite	234	338	4.3 ms	3.0 ms	10.7 ms	8.4 ms	GitHub/PINTO	DFP
203	MediaPipe-selfie-segmentation	256×256×3	103.8 K	TFLite	691	934	1.4 ms	1.1 ms	4.0 ms	3.3 ms	MediaPipe	DFP
204	MediaPipe-selfie-segmentation-landscape	144×256×3	103.9 K	TFLite	1415	2001	0.7 ms	0.5 ms	2.7 ms	2.2 ms	MediaPipe	DFP
205	PaNet-ResNet-18-ffm	640×640×3	11.7 M	ONNX	27	39	37.0 ms	25.6 ms	56.9 ms	40.4 ms	OpenMMLab	DFP
206	STDC1-CityScapes	512×512×3	7.9 M	ONNX	156	221	6.4 ms	4.5 ms	12.3 ms	9.2 ms	OpenMMLab	DFP
207	STDC2-CityScapes	512×512×3	11.7 M	ONNX	144	204	6.9 ms	4.9 ms	18.0 ms	12.8 ms	OpenMMLab	DFP
208	Selfie-Segmentation	256×256×3	103.9 K	TFLite	766	1070	1.3 ms	0.9 ms	4.5 ms	3.5 ms	GitHub/PINTO	DFP
209	U-NetMobileNet-v2	256×256×3	6.2 M	TFLite	51	72	19.6 ms	13.9 ms	23.5 ms	17.0 ms	Coral	DFP
210	U-NetMobileNet-v2	128×128×3	6.2 M	TFLite	202	286	5.0 ms	3.5 ms	7.0 ms	5.2 ms	Coral	DFP
211	UNet	1024×1024×3	1.9 M	TFLite	27	32	37.0 ms	31.2 ms	42.4 ms	35.9 ms	AI Benchmark	DFP
212	UNet	256×256×1	29.6 M	ONNX	31	44	32.3 ms	22.7 ms	35.5 ms	25.3 ms	Zetane Viewer	DFP
213	Xray-80-whole-f2	256×256×1	29.6 M	Keras	34	48	29.4 ms	20.8 ms	43.5 ms	30.9 ms	Zetane Viewer	DFP
214	YOLO-v8-medium-seg	640×640×3	26.0 M	ONNX	27	38	37.0 ms	26.3 ms	60.4 ms	42.7 ms	Ultralytics	DFP
215	YOLO-v8-nano-seg	640×640×3	3.2 M	ONNX	209	215	4.8 ms	4.7 ms	11.7 ms	10.8 ms	Ultralytics	DFP

Pose Estimation

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
216	BodyPix (MobileNet)	480×640×3	1.2 M	TFLite	413	497	2.4 ms	2.0 ms	4.2 ms	3.6 ms	Coral	DFP
217	BodyPix (MobileNet)	512×512×3	1.2 M	TFLite	475	574	2.1 ms	1.7 ms	3.8 ms	3.2 ms	Coral	DFP
218	BodyPix (MobileNet)	768×1024×3	1.2 M	TFLite	170	204	5.9 ms	4.9 ms	9.5 ms	8.0 ms	Coral	DFP
219	BodyPix (MobileNet)	576×768×3	1.2 M	TFLite	297	357	3.4 ms	2.8 ms	5.9 ms	4.9 ms	Coral	DFP
220	BodyPix (MobileNet)	720×1280×3	1.2 M	TFLite	146	176	6.8 ms	5.7 ms	11.7 ms	9.5 ms	Coral	DFP
221	BodyPix (MobileNet)	352×480×3	1.2 M	TFLite	804	970	1.2 ms	1.0 ms	3.0 ms	2.4 ms	Coral	DFP
222	BodyPix (ResNet-50)	496×768×3	22.7 M	TFLite	152	215	6.6 ms	4.7 ms	17.3 ms	12.6 ms	Coral	DFP
223	BodyPix (ResNet-50)	624×864×3	22.7 M	TFLite	112	159	8.9 ms	6.3 ms	20.1 ms	14.7 ms	Coral	DFP
224	BodyPix (ResNet-50)	288×416×3	22.7 M	TFLite	247	349	4.0 ms	2.9 ms	12.2 ms	8.9 ms	Coral	DFP
225	BodyPix (ResNet-50)	480×640×3	22.7 M	TFLite	100	139	10.0 ms	7.2 ms	23.5 ms	17.2 ms	Coral	DFP
226	BodyPix (ResNet-50)	672×928×3	22.7 M	TFLite	53	76	18.9 ms	13.2 ms	35.3 ms	25.1 ms	Coral	DFP
227	HHP-Net	1×5×1, 1×5×1, 1×5×1	36.4 K	TFLite	11657	11716	0.1 ms	0.1 ms	0.2 ms	0.2 ms	GitHub/PINTO	DFP
228	Handpose-estimation-mediapipe	224×224×3	988.0 K	ONNX	1937	2304	0.5 ms	0.4 ms	1.4 ms	1.1 ms	OpenCV	DFP
229	Handpose-estimation-mediapipe8	224×224×3	904.1 K	ONNX	425	602	2.4 ms	1.7 ms	6.0 ms	4.4 ms	OpenCV	DFP
230	Human-pose-estimation-3d-0001	256×448×3	4.8 M	ONNX	461	652	2.2 ms	1.5 ms	8.8 ms	6.5 ms	OpenVINO	DFP
231	MediaPipe-pose-landk	256×256×3	3.0 M	TFLite	748	880	1.3 ms	1.1 ms	4.3 ms	3.5 ms	MediaPipe	DFP
232	MediaPipe-pose-landk-heavy	256×256×3	13.1 M	TFLite	103	146	9.7 ms	6.8 ms	14.4 ms	10.5 ms	MediaPipe	DFP
233	MediaPipe-pose-landk-lite	256×256×3	1.3 M	TFLite	716	883	1.4 ms	1.1 ms	3.6 ms	3.0 ms	MediaPipe	DFP
234	MoveNet.SinglePose.Lightning	192×192×3	2.2 M	TFLite	2542	2730	0.4 ms	0.4 ms	2.1 ms	1.8 ms	Coral	DFP
235	MoveNet.SinglePose.Thunder	256×256×3	5.9 M	TFLite	1417	1465	0.7 ms	0.7 ms	4.0 ms	3.2 ms	Coral	DFP
236	Pose-estimation-mediapipe	258×258×3	1.3 M	ONNX	718	906	1.4 ms	1.1 ms	3.6 ms	3.0 ms	OpenCV	DFP
237	PoseC3D	48×64×64×17	2.0 M	ONNX	29	41	34.5 ms	24.4 ms	41.7 ms	30.3 ms	GitHub/PINTO	DFP
238	PoseC3D	48×56×56×17	2.9 M	ONNX	33	47	30.3 ms	21.3 ms	36.8 ms	26.5 ms	GitHub/PINTO	DFP
239	PoseNet (MobileNet)	353×481×3	1.2 M	TFLite	802	968	1.2 ms	1.0 ms	2.6 ms	2.2 ms	Coral	DFP
240	PoseNet (MobileNet)	481×641×3	1.2 M	TFLite	413	502	2.4 ms	2.0 ms	4.2 ms	3.5 ms	Coral	DFP
241	PoseNet (MobileNet)	721×1281×3	1.2 M	TFLite	146	177	6.8 ms	5.6 ms	10.8 ms	9.1 ms	Coral	DFP
242	PoseNet (ResNet-50)	288×416×3	22.6 M	TFLite	258	365	3.9 ms	2.7 ms	12.1 ms	8.8 ms	Coral	DFP
243	PoseNet (ResNet-50)	496×768×3	22.6 M	TFLite	153	217	6.5 ms	4.6 ms	16.5 ms	12.0 ms	Coral	DFP
244	PoseNet (ResNet-50)	736×960×3	22.6 M	TFLite	83	117	12.0 ms	8.5 ms	24.7 ms	17.9 ms	Coral	DFP
245	PoseNet (ResNet-50)	624×864×3	22.6 M	TFLite	112	159	8.9 ms	6.3 ms	20.2 ms	14.7 ms	Coral	DFP
246	PoseNet (ResNet-50)	480×640×3	22.6 M	TFLite	117	165	8.5 ms	6.1 ms	20.4 ms	14.6 ms	Coral	DFP
247	PoseNet (ResNet-50)	672×928×3	22.6 M	TFLite	53	75	18.9 ms	13.3 ms	35.4 ms	25.2 ms	Coral	DFP
248	RtmDet-nano-person	320×320×3	962.8 K	ONNX	1194	1443	0.8 ms	0.7 ms	3.4 ms	2.6 ms	OpenMMLab	DFP
249	YOLO-v8-medium-pose	640×640×3	25.2 M	ONNX	43	62	23.3 ms	16.1 ms	38.3 ms	27.3 ms	Ultralytics	DFP
250	YOLO-v8-nano-pose	640×640×3	3.1 M	ONNX	327	389	3.1 ms	2.6 ms	7.6 ms	6.6 ms	Ultralytics	DFP
251	YOLO-v8-small-pose	640×640×3	11.1 M	ONNX	228	323	4.4 ms	3.1 ms	11.5 ms	8.6 ms	Ultralytics	DFP

Image Enhancement/Super Resolution

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
252	AOD-Net	720×1280×3	1.7 K	TFLite	17	24	58.8 ms	41.7 ms	62.6 ms	44.9 ms	GitHub/PINTO	DFP
253	AOD-Net	480×640×3	1.7 K	TFLite	52	73	19.2 ms	13.7 ms	21.8 ms	15.4 ms	GitHub/PINTO	DFP
254	Colorization-v2	256×256×1	30.7 M	ONNX	36	51	27.8 ms	19.6 ms	40.7 ms	29.1 ms	OpenVINO	DFP
255	DPED	384×512×48	414.7 K	TFLite	41	45	24.4 ms	22.2 ms	49.7 ms	47.9 ms	AI Benchmark	DFP
256	ESRGAN	128×128×3	1.2 M	TFLite	61	87	16.4 ms	11.5 ms	23.9 ms	17.1 ms	AI Benchmark	DFP
257	Fast-SRGAN	512×512×3	154.9 K	TFLite	4	5	250.0 ms	200.0 ms	272.3 ms	195.5 ms	GitHub/PINTO	DFP
258	Fast-SRGAN	128×128×3	154.9 K	TFLite	62	87	16.1 ms	11.5 ms	17.4 ms	12.6 ms	GitHub/PINTO	DFP
259	IMDN	480×640×3	873.0 K	TFLite	9	12	111.1 ms	83.3 ms	145.4 ms	110.8 ms	GitHub/PINTO	DFP
260	IMDN	360×640×3	873.0 K	TFLite	12	17	83.3 ms	58.8 ms	111.8 ms	85.1 ms	GitHub/PINTO	DFP
261	IMDN	1024×1024×3	149.1 K	TFLite	40	42	25.0 ms	23.8 ms	30.2 ms	28.4 ms	AI Benchmark	DFP
262	MBLLEN	480×640×3	439.6 K	TFLite	11	15	90.9 ms	66.7 ms	99.5 ms	70.3 ms	GitHub/PINTO	DFP
263	MBLLEN	720×1280×3	439.6 K	TFLite	4	5	250.0 ms	200.0 ms	291.2 ms	206.1 ms	GitHub/PINTO	DFP
264	Noise2Noise	256×256×3	1.2 M	TFLite	61	86	16.4 ms	11.6 ms	20.8 ms	15.0 ms	GitHub/PINTO	DFP
265	PMN	416×416×4	7.4 M	ONNX	43	61	23.3 ms	16.4 ms	27.4 ms	19.6 ms	GitHub/PINTO	DFP
266	SCI	288×800×3	0.2 K	ONNX	426	557	2.3 ms	1.8 ms	3.7 ms	3.2 ms	GitHub/PINTO	DFP
267	SCL-LLE	240×320×3	77.6 K	TFLite	104	148	9.6 ms	6.8 ms	13.5 ms	9.8 ms	GitHub/PINTO	DFP
268	SCL-LLE	180×320×3	77.6 K	TFLite	138	195	7.2 ms	5.1 ms	10.2 ms	7.6 ms	GitHub/PINTO	DFP
269	SRCNN-di-v2-k	896×896×3	19.6 K	ONNX	7	10	142.9 ms	100.0 ms	139.0 ms	98.8 ms	OpenMMLab	DFP
270	SRGAN	256×256×3	1.4 M	TFLite	8	11	125.0 ms	90.9 ms	137.3 ms	97.2 ms	AI Benchmark	DFP
271	Semantic-Guided-Low-Light-Image-Enhanceme…	24×54×3	10.3 K	ONNX	6654	6969	0.2 ms	0.1 ms	0.4 ms	0.3 ms	GitHub/PINTO	DFP
272	Super-Resolution	224×224×1	58.3 K	ONNX	160	227	6.2 ms	4.4 ms	7.0 ms	5.0 ms	ONNX Zoo	DFP
273	TBEFN	180×320×3	475.0 K	TFLite	199	226	5.0 ms	4.4 ms	6.2 ms	5.5 ms	GitHub/PINTO	DFP
274	TBEFN	240×320×3	475.0 K	TFLite	158	178	6.3 ms	5.6 ms	7.7 ms	6.9 ms	GitHub/PINTO	DFP
275	VSR	540×960×3	18.6 K	TFLite	28	32	35.7 ms	31.2 ms	88.3 ms	77.3 ms	AI Benchmark	DFP
276	XLSR	360×640×3	21.5 K	TFLite	70	99	14.3 ms	10.1 ms	23.2 ms	19.2 ms	AI Benchmark	DFP
277	Y-Net	480×640×3	190.9 K	TFLite	33	40	30.3 ms	25.0 ms	32.9 ms	26.5 ms	GitHub/PINTO	DFP
278	Y-Net	240×320×3	190.9 K	TFLite	130	161	7.7 ms	6.2 ms	8.8 ms	7.1 ms	GitHub/PINTO	DFP
279	Zero-DCE	360×640×3	77.6 K	TFLite	40	42	25.0 ms	23.8 ms	33.1 ms	31.9 ms	GitHub/PINTO	DFP
280	Zero-DCE	240×320×3	77.6 K	TFLite	121	126	8.3 ms	7.9 ms	12.2 ms	11.0 ms	GitHub/PINTO	DFP
281	Zero-DCE-improved	360×512×3	23.0 K	ONNX	87	123	11.5 ms	8.1 ms	19.4 ms	15.1 ms	GitHub/PINTO	DFP

Face/Head

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
282	ACR-Loss	224×224×3	2.3 M	TFLite	1937	2310	0.5 ms	0.4 ms	1.5 ms	1.2 ms	GitHub/PINTO	DFP
283	BlazeFace	128×128×3	99.0 K	TensorFlow	6811	7834	0.1 ms	0.1 ms	0.5 ms	0.5 ms	GitHub/PINTO	DFP
284	Face-Detection-0100	256×256×3	1.7 M	TFLite	1676	2015	0.6 ms	0.5 ms	1.7 ms	1.3 ms	GitHub/PINTO	DFP
285	Face-Detection-YuNet	640×640×3	51.9 K	ONNX	327	398	3.1 ms	2.5 ms	4.6 ms	4.0 ms	OpenCV	DFP
286	Face-Detection-adas-0001	384×672×3	1.0 M	TFLite	531	649	1.9 ms	1.5 ms	3.6 ms	2.9 ms	GitHub/PINTO	DFP
287	Face-Mask-Detection	300×300×3	2.5 M	TFLite	840	1189	1.2 ms	0.8 ms	2.8 ms	2.2 ms	GitHub/PINTO	DFP
288	Face-landk	160×160×3	611.3 K	ONNX	4238	4896	0.2 ms	0.2 ms	0.8 ms	0.6 ms	GitHub/PINTO	DFP
289	Face-recognition-sface	112×112×3	9.2 M	ONNX	2580	3652	0.4 ms	0.3 ms	1.4 ms	1.1 ms	OpenCV	DFP
290	FaceBoxes	1024×1024×3	982.3 K	ONNX	128	156	7.8 ms	6.4 ms	10.8 ms	9.0 ms	OpenVINO	DFP
291	FaceMesh	192×192×3	590.5 K	TFLite	3591	4220	0.3 ms	0.2 ms	0.8 ms	0.7 ms	GitHub/PINTO	DFP
292	License-plate-Detection-LPD-YuNet	240×320×3	1008.7 K	ONNX	726	913	1.4 ms	1.1 ms	2.4 ms	2.0 ms	OpenCV	DFP
293	MediaPipe-face-Detection-range	192×192×3	502.1 K	TFLite	2583	3516	0.4 ms	0.3 ms	1.5 ms	1.3 ms	MediaPipe	DFP
294	MediaPipe-face-Detection-short-range	128×128×3	99.1 K	TFLite	6809	7838	0.1 ms	0.1 ms	0.5 ms	0.5 ms	MediaPipe	DFP
295	MediaPipe-iris-landk	64×64×3	636.0 K	TFLite	10531	10531	0.1 ms	0.1 ms	0.6 ms	0.6 ms	MediaPipe	DFP
296	RetinaFacewithResNet-50	640×640×3	26.0 M	ONNX	45	64	22.2 ms	15.6 ms	34.6 ms	24.5 ms	OpenVINO	DFP
297	SFace	112×112×3	9.2 M	TFLite	2579	3650	0.4 ms	0.3 ms	1.4 ms	1.1 ms	GitHub/PINTO	DFP
298	SSD-MobileNet-v2 (Face)	320×320×3	5.1 M	TFLite	1195	1437	0.8 ms	0.7 ms	2.6 ms	2.0 ms	Coral	DFP
299	SynergyNet	120×120×3	2.2 M	ONNX	4944	6984	0.2 ms	0.1 ms	0.9 ms	0.8 ms	GitHub/PINTO	DFP
300	UltraLightweightFaceDetectionRFB320	240×320×3	274.5 K	ONNX	1685	2011	0.6 ms	0.5 ms	1.3 ms	1.1 ms	OpenVINO	DFP
301	UltraLightweightFaceDetectionslim320	240×320×3	258.5 K	ONNX	1685	2014	0.6 ms	0.5 ms	1.3 ms	1.1 ms	OpenVINO	DFP
302	Ultraface	480×640×3	274.5 K	ONNX	414	502	2.4 ms	2.0 ms	3.6 ms	3.2 ms	ONNX Zoo	DFP
303	Version-RFB	240×320×3	274.5 K	ONNX	1681	2012	0.6 ms	0.5 ms	1.4 ms	1.1 ms	ONNX Zoo	DFP
304	WHENet	224×224×3	4.1 M	TFLite	971	1374	1.0 ms	0.7 ms	6.4 ms	4.8 ms	GitHub/PINTO	DFP
305	YOLO-v5-Face	256×320×3	1.6 M	TensorFlow	1635	1983	0.6 ms	0.5 ms	2.8 ms	2.2 ms	GitHub/PINTO	DFP
306	YuNet	120×160×3	81.1 K	TensorFlow	5921	6849	0.2 ms	0.1 ms	0.5 ms	0.5 ms	GitHub/PINTO	DFP
307	YuNet-v2	640×640×3	51.9 K	TFLite	327	398	3.1 ms	2.5 ms	4.8 ms	3.9 ms	GitHub/PINTO	DFP

Classification (Other)

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
308	Aiy-vision-classifier-birds	224×224×3	3.3 M	TFLite	1785	2299	0.6 ms	0.4 ms	2.0 ms	1.6 ms	TFHub	DFP
309	Aiy-vision-classifier-food	192×192×3	5.0 M	TFLite	3578	4212	0.3 ms	0.2 ms	1.4 ms	1.1 ms	TFHub	DFP
310	Aiy-vision-classifier-insects	224×224×3	3.4 M	TFLite	1785	2299	0.6 ms	0.4 ms	1.9 ms	1.6 ms	TFHub	DFP
311	Aiy-vision-classifier-plants	224×224×3	4.7 M	TFLite	1785	2308	0.6 ms	0.4 ms	2.0 ms	1.6 ms	TFHub	DFP
312	Anti-spoof-mn3	128×128×3	2.9 M	TFLite	4381	6190	0.2 ms	0.2 ms	1.7 ms	1.3 ms	GitHub/PINTO	DFP
313	Gender-googleNet	224×224×3	5.7 M	ONNX	519	734	1.9 ms	1.4 ms	3.8 ms	2.8 ms	ONNX Zoo	DFP
314	Malaria-Detector	125×125×3	6.7 M	Keras	6801	8037	0.1 ms	0.1 ms	0.5 ms	0.5 ms	xmartlabs	DFP
315	NSFW	224×224×3	5.6 M	TFLite	1829	2221	0.5 ms	0.5 ms	2.0 ms	1.6 ms	GitHub/PINTO	DFP
316	Person-Attributes-recog-crossroad-0230	160×80×3	717.7 K	TFLite	5821	8058	0.2 ms	0.1 ms	0.6 ms	0.5 ms	GitHub/PINTO	DFP
317	Vehicle-Attributes-recog-barrier-0039	72×72×3	611.5 K	TFLite	7148	8714	0.1 ms	0.1 ms	0.6 ms	0.5 ms	GitHub/PINTO	DFP
318	Vehicle-Attributes-recog-barrier-0042	72×72×3	10.7 M	TFLite	5910	8367	0.2 ms	0.1 ms	1.2 ms	0.9 ms	GitHub/PINTO	DFP

ADAS/Driver Monitoring

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
319	Driver-action-recognition-adas-0002-encod…	224×224×3	2.7 M	TFLite	1926	2294	0.5 ms	0.4 ms	1.6 ms	1.3 ms	GitHub/PINTO	DFP
320	Gaze-estimation-adas-0002	60×60×3, 60×60×3, 3	1.8 M	TFLite	258	260	3.9 ms	3.8 ms	4.5 ms	4.4 ms	GitHub/PINTO	DFP
321	Head-pose-estimation-adas-0001	60×60×3	1.8 M	TFLite	8393	11572	0.1 ms	0.1 ms	0.7 ms	0.6 ms	GitHub/PINTO	DFP
322	Open-closed-eye-0001	32×32×3	11.0 K	ONNX	19354	19354	0.1 ms	0.1 ms	0.2 ms	0.2 ms	OpenVINO	DFP
323	Pedestrian-Detection-adas-0002	384×672×3	1.1 M	TFLite	531	645	1.9 ms	1.6 ms	3.6 ms	3.1 ms	GitHub/PINTO	DFP
324	Pedestrian-and-vehicle-Detector-adas-0001	384×672×3	1.6 M	TFLite	531	645	1.9 ms	1.6 ms	4.0 ms	3.3 ms	GitHub/PINTO	DFP
325	Road-segmentation-adas-0001	512×896×3	180.0 K	TFLite	145	186	6.9 ms	5.4 ms	10.3 ms	7.7 ms	GitHub/PINTO	DFP
326	Vehicle-Detection-0200	256×256×3	1.7 M	TFLite	1677	2014	0.6 ms	0.5 ms	1.7 ms	1.4 ms	GitHub/PINTO	DFP
327	YOLOp320	320×320×3	7.6 M	ONNX	403	570	2.5 ms	1.8 ms	7.2 ms	5.3 ms	GitHub/YOLOP	DFP
328	YOLOp640	640×640×3	7.6 M	ONNX	134	190	7.5 ms	5.3 ms	19.1 ms	13.6 ms	GitHub/YOLOP	DFP

Hand/Palm

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
329	Hand-Detection-and-Tracking	256×256×3	1.3 M	TensorFlow	1035	1126	1.0 ms	0.9 ms	4.3 ms	3.4 ms	GitHub/PINTO	DFP
330	Hand-recrop	256×256×3	26.1 K	TFLite	1683	2028	0.6 ms	0.5 ms	1.0 ms	0.8 ms	GitHub/PINTO	DFP
331	MediaPipe-hand-landk	224×224×3	2.6 M	TFLite	1946	2299	0.5 ms	0.4 ms	1.9 ms	1.5 ms	MediaPipe	DFP
332	MediaPipe-hand-landk-lite	224×224×3	988.0 K	TFLite	1943	2320	0.5 ms	0.4 ms	1.4 ms	1.2 ms	MediaPipe	DFP
333	MediaPipe-palm-Detection	192×192×3	1.1 M	TFLite	907	1283	1.1 ms	0.8 ms	2.6 ms	2.0 ms	MediaPipe	DFP
334	MediaPipe-palm-Detection-lite	192×192×3	937.6 K	TFLite	1259	1737	0.8 ms	0.6 ms	2.2 ms	1.7 ms	MediaPipe	DFP
335	Palm-Detection-mediapipe	192×192×3	937.6 K	ONNX	1145	1553	0.9 ms	0.6 ms	2.2 ms	1.7 ms	OpenCV	DFP

Depth Estimation

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
336	Graft-PSMNet	120×160×256	1.2 M	ONNX	96	96	10.4 ms	10.4 ms	19.6 ms	19.6 ms	GitHub/PINTO	DFP
337	Graft-PSMNet	80×120×256	1.2 M	ONNX	193	193	5.2 ms	5.2 ms	9.3 ms	9.3 ms	GitHub/PINTO	DFP
338	HR-Depth	384×1280×3	10.7 M	ONNX	54	55	18.5 ms	18.2 ms	38.3 ms	39.6 ms	GitHub/PINTO	DFP
339	PyDNet	192×320×3	1.9 M	TFLite	354	501	2.8 ms	2.0 ms	5.0 ms	3.8 ms	GitHub/PINTO	DFP
340	PyDNet2	384×1280×3	699.9 K	TFLite	44	63	22.7 ms	15.9 ms	28.2 ms	20.9 ms	GitHub/PINTO	DFP
341	PyDNet2	384×640×3	699.9 K	TFLite	89	126	11.2 ms	7.9 ms	15.1 ms	10.8 ms	GitHub/PINTO	DFP

Style Transfer/GANs

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
342	Arbitrary-stylization	256×256×3	2.2 M	TFLite	1680	2015	0.6 ms	0.5 ms	1.6 ms	1.4 ms	GitHub/PINTO	DFP
343	Arbitrary-stylization-inception-v3	256×256×3	8.6 M	TFLite	566	802	1.8 ms	1.2 ms	4.8 ms	3.5 ms	TFHub	DFP
344	Facial-cartoonizer	512×512×3	2.0 M	ONNX	62	87	16.1 ms	11.5 ms	19.5 ms	14.1 ms	GitHub/Facial Image Cartoonizaiton	DFP
345	Magenta-arbitrary-stylization-256	256×256×3	2.2 M	TFLite	1679	2015	0.6 ms	0.5 ms	1.7 ms	1.4 ms	TFHub	DFP
346	MobileStyleGAN	512	2.0 M	ONNX	33179	33179	0.0 ms	0.0 ms	0.3 ms	0.3 ms	GitHub/PINTO	DFP

Emotion Detection

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
347	Emotion-FERPlus	64×64×1	8.4 M	TFLite	965	1367	1.0 ms	0.7 ms	1.8 ms	1.3 ms	GitHub/PINTO	DFP
348	Emotion-ferplus	64×64×1	8.4 M	ONNX	991	1404	1.0 ms	0.7 ms	1.7 ms	1.3 ms	ONNX Zoo	DFP
349	Facial-expRession-recognition-MobilefaceN…	112×112×3	1.1 M	TFLite	5722	7626	0.2 ms	0.1 ms	0.9 ms	0.8 ms	GitHub/PINTO	DFP
350	Facial-expRession-recognition-MobilefaceN…	112×112×3	1.1 M	ONNX	5719	7619	0.2 ms	0.1 ms	0.9 ms	0.8 ms	OpenCV	DFP

Audio/Speech

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
351	FRILL	96×64×1	9.6 M	TFLite	5469	7733	0.2 ms	0.1 ms	1.3 ms	1.0 ms	GitHub/PINTO	DFP
352	Light-SERNet	184×40×1	449.9 K	TFLite	3221	3364	0.3 ms	0.3 ms	0.7 ms	0.6 ms	GitHub/PINTO	DFP
353	Speech-enhancement	128×128×1	1.9 M	ONNX	968	1370	1.0 ms	0.7 ms	2.1 ms	1.6 ms	GitHub/PINTO	DFP
354	YAMNet	96×64×1	3.6 M	TFLite	11435	11862	0.1 ms	0.1 ms	0.8 ms	0.7 ms	GitHub/PINTO	DFP

Gesture Recognition

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
355	American-sign-language	224×224×3	2.7 M	TFLite	1933	2310	0.5 ms	0.4 ms	1.5 ms	1.2 ms	TFHub	DFP
356	FINNger	96×96×3	3.2 M	TensorFlow	379	537	2.6 ms	1.9 ms	3.8 ms	2.8 ms	GitHub/PINTO	DFP
357	Naruto-handsign-Detection	416×416×3	871.1 K	TFLite	809	845	1.2 ms	1.2 ms	3.3 ms	2.9 ms	GitHub/PINTO	DFP

Tracking

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
358	LiteTrack	127×127×3	1.2 M	TFLite	2447	3462	0.4 ms	0.3 ms	2.6 ms	2.0 ms	GitHub/PINTO	DFP
359	OpenCV-DaSiamRPN (Feature Extractor)	360×480×3	4.8 M	ONNX	93	132	10.8 ms	7.6 ms	12.5 ms	9.1 ms	OpenCV	DFP
360	OpenCV-DaSiamRPN (Kernel-R1)	24×24×256	11.3 M	ONNX	111	115	9.0 ms	8.7 ms	13.4 ms	12.6 ms	OpenCV	DFP

Point Cloud

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
361	PointPillars (SECFPN) - KITTI 3D	496×432×64	4.6 M	ONNX	10	15	100.0 ms	66.7 ms	116.7 ms	88.8 ms	OpenMMLab	DFP
362	PointPillars (SECFPN-SBN) - NUS 3D	400×400×64	4.7 M	ONNX	13	18	76.9 ms	55.6 ms	113.0 ms	90.7 ms	OpenMMLab	DFP
363	SFA3D	608×608×3	12.1 M	TensorFlow	90	95	11.1 ms	10.5 ms	19.4 ms	16.7 ms	GitHub/PINTO	DFP

Wireframe/keypoint Detection

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
364	M-LSD	320×320×4	599.8 K	TFLite	804	1090	1.2 ms	0.9 ms	3.0 ms	2.4 ms	GitHub/PINTO	DFP
365	M-LSD	512×512×4	1.5 M	TFLite	31	44	32.3 ms	22.7 ms	37.3 ms	27.0 ms	GitHub/PINTO	DFP

Re-identification

#	Name	Input Resolution	Params Count	F/W	FPS (14T)	FPS (20T)	Latency 1 (14T)	Latency 1 (20T)	Latency 2 (14T)	Latency 2 (20T)	Source	DFP
366	Person-ReID-YouTube	256×128×3	25.5 M	ONNX	736	1042	1.4 ms	1.0 ms	4.1 ms	3.0 ms	OpenCV	DFP

Third-Party Licenses

Each of these models is shared under the original creator’s license. Please check the license at the source page.