tflite_models_checklist.md

Model validation and performance analysis status for TensorFlow Lite

Public models (Open Model Zoo)

Image classification

Model	Availability in OMZ (2023.03.10)	Availability in the validation table
densenet-121	+-	+
googlenet-v1-tf	+-	+
googlenet-v2-tf	+-	+
googlenet-v3	+-	+
googlenet-v4-tf	+-	+
mobilenet-v1-1.0-224-tf	+-	+
mobilenet-v2-1.0-224	+-	+
mobilenet-v2-1.4-224	+-	+
mobilenet-v3-small-1.0-224	+-	+
mobilenet-v3-large-1.0-224	+-	+
inception-resnet-v2-tf	+	+
resnet-50-tf	+-	+

Notes:

1. Inference implementation for several GoogleNet-models supports batch size that equals 1.
2. "+-" in the column of availability in OMZ (2023.03.10) means that model was converted by omz_converter from .ckpt into .pb if it is required, and by internal converter src/model_converters/tflite_converter.py from .pb into .tflite.

Semantic segmentation

Model	Availability in OMZ (2023.09.29)	Availability in the validation table
deeplabv3	+	+

Other tasks

[TBD]

Public models (TF hub)

Image classification

Model	Availability in TF hub (2023.03.10)	Availability in the validation table
lite-model_mobilenet_v1_100_224_fp32_1.tflite	+ (link)	+
lite-model_mobilenet_v2_100_224_fp32_1.tflite	+ (link)	+
lite-model_mobilenet_v3_small_100_224_fp32_1.tflite	+ (link)	+
lite-model_mobilenet_v3_large_100_224_fp32_1.tflite	+ (link)	+
lite-model_mobilenet_v1_100_224_uint8_1.tflite	+ (link)	+
lite-model_mobilenet_v2_100_224_uint8_1.tflite	+ (link)	+
lite-model_mobilenet_v3_small_100_224_uint8_1.tflite	+ (link)	+
lite-model_mobilenet_v3_large_100_224_uint8_1.tflite	+ (link)	+
efficientnet_lite0_fp32_2.tflite	+ (link)	+
efficientnet_lite1_fp32_2.tflite	+ (link)	+
efficientnet_lite2_fp32_2.tflite	+ (link)	+
efficientnet_lite3_fp32_2.tflite	+ (link)	+
efficientnet_lite4_fp32_2.tflite	+ (link)	+
lite-efficientnet_lite0_uint8_2.tflite	+ (link)	+
lite-efficientnet_lite1_uint8_2.tflite	+ (link)	+
lite-efficientnet_lite2_uint8_2.tflite	+ (link)	+
lite-efficientnet_lite3_uint8_2.tflite	+ (link)	+
lite-efficientnet_lite4_uint8_2.tflite	+ (link)	+
efficientnet_lite0_int8_2.tflite	+ (link)	+
efficientnet_lite1_int8_2.tflite	+ (link)	+
efficientnet_lite2_int8_2.tflite	+ (link)	+
efficientnet_lite3_int8_2.tflite	+ (link)	+
efficientnet_lite4_int8_2.tflite	+ (link)	+

Note: inference implementation for EfficientNet-models supported for batch size that equals 1.

Other tasks

[TBD]