Blame - NnapiSupport.txt - ml/android-nn-driver

2018-03-09 13:51:08 +0000

[diff] [blame]

2

Matthew Bentham

a3e23ca

2019-05-13 12:49:59 +0100

[diff] [blame]

3

This release of ArmNN for Android supports use as a driver for the Android Neural Networks API. It implements the

Aron Virginas-Tar

2019-06-17 15:48:45 +0100

[diff] [blame]

4

android.hardware.neuralnetworks@1.0, android.hardware.neuralnetworks@1.1 and android.hardware.neuralnetworks@1.2

5

HAL interfaces.

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

6

7

For more information on the Android Neural Networks API, see https://developer.android.com/ndk/guides/neuralnetworks/index.html

8

9

For integration and usage documentation, please see README.md.

10

11

--- Support for Android Neural Networks HAL operations ---

12

Aron Virginas-Tar

2019-06-17 15:48:45 +0100

[diff] [blame]

13

The following AndroidNN HAL 1.0 and 1.1 operations are currently supported:

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

14

15

AndroidNN operator Tensor type supported

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

16

ADD (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

17

AVERAGE_POOL_2D (FLOAT32,QUANT8_ASYMM)

Sadik Armagan

82fbeb3

2018-12-20 17:12:26 +0000

[diff] [blame]

18

BATCH_TO_SPACE_ND (FLOAT32,QUANT8_ASYMM)

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

19

CONCATENATION (FLOAT32,QUANT8_ASYMM)

20

CONV_2D (FLOAT32,QUANT8_ASYMM)

21

DEPTHWISE_CONV_2D (FLOAT32,QUANT8_ASYMM)

22

DIV (FLOAT32,QUANT8_ASYMM)

23

DEQUANTIZE (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

24

FLOOR (FLOAT32)

Matthew Bentham

378333d

2018-10-01 16:07:27 +0100

[diff] [blame]

25

FULLY_CONNECTED (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

26

L2_NORMALIZATION (FLOAT32)

nikraj01

c6494cc

2019-05-22 16:30:44 +0100

[diff] [blame]

27

L2_POOL_2D (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

28

LOCAL_RESPONSE_NORMALIZATION (FLOAT32)

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

29

LOGISTIC (FLOAT32,QUANT8_ASYMM)

Ferran Balaguer

d04c043

2018-11-15 14:48:05 +0000

[diff] [blame]

30

LSTM (FLOAT32)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

31

MAX_POOL_2D (FLOAT32,QUANT8_ASYMM)

Ferran Balaguer

d04c043

2018-11-15 14:48:05 +0000

[diff] [blame]

32

MEAN (FLOAT32,QUANT8_ASYMM)

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

33

MUL (FLOAT32,QUANT8_ASYMM)

Ferran Balaguer

d04c043

2018-11-15 14:48:05 +0000

[diff] [blame]

34

PAD (FLOAT32,QUANT8_ASYMM)

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

35

RELU (FLOAT32,QUANT8_ASYMM)

36

RELU1 (FLOAT32,QUANT8_ASYMM)

37

RELU6 (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

38

RESHAPE (FLOAT32,QUANT8_ASYMM)

nikraj01

c6494cc

2019-05-22 16:30:44 +0100

[diff] [blame]

39

RESIZE_BILINEAR (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

40

SOFTMAX (FLOAT32,QUANT8_ASYMM)

Sadik Armagan

82fbeb3

2018-12-20 17:12:26 +0000

[diff] [blame]

41

SPACE_TO_BATCH_ND (FLOAT32,QUANT8_ASYMM)

Keith Davis

a6bc52f

2019-06-26 09:39:49 +0100

[diff] [blame]

42

SPACE_TO_DEPTH_ND (FLOAT32,QUANT8_ASYMM)

saoste01

b847148

2018-10-10 09:44:51 +0100

[diff] [blame]

43

SQUEEZE (FLOAT32,QUANT8_ASYMM)

Sadik Armagan

82fbeb3

2018-12-20 17:12:26 +0000

[diff] [blame]

44

STRIDED_SLICE (FLOAT32,QUANT8_ASYMM)

Matthew Bentham

2019-06-05 09:24:31 +0100

[diff] [blame]

45

SUB (FLOAT32,QUANT8_ASYMM)

46

TANH (FLOAT32,QUANT8_ASYMM)

saoste01

fe46315

2018-10-18 17:49:56 +0100

[diff] [blame]

47

TRANSPOSE (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

48

Aron Virginas-Tar

2019-06-17 15:48:45 +0100

[diff] [blame]

49

The following AndroidNN HAL 1.2 operations are currently supported:

50

51

CONV_2D (FLOAT32,QUANT8_ASYMM)

52

DEPTHWISE_CONV_2D (FLOAT32,QUANT8_ASYMM)

Narumol Prangnawarat

95b1ef6

2019-07-15 12:02:20 +0100

[diff] [blame]

53

MAXIMUM (FLOAT32,QUANT8_ASYMM)

Ellen Norris-Thompson

1cb29aa

2019-07-11 17:27:37 +0100

[diff] [blame]

54

MINIMUM (FLOAT32,QUANT8_ASYMM)

Sadik Armagan

5a476a8

2019-07-30 09:43:18 +0100

[diff] [blame]

55

PAD_V2 (FLOAT32,QUANT8_ASYMM)

Matteo Martincigh

0bd89a8

2019-07-02 16:53:10 +0100

[diff] [blame]

56

PRELU (FLOAT32,QUANT8_ASYMM)

Sadik Armagan

5a476a8

2019-07-30 09:43:18 +0100

[diff] [blame]

57

QUANTIZE (FLOAT32,QUANT8_ASYMM)

Ellen Norris-Thompson

7efb46d

2019-07-24 17:39:19 +0100

[diff] [blame]

58

QUANTIZED_16BIT_LSTM (QUANT8_ASYMM)

Aron Virginas-Tar

7a6d11b

2019-07-03 15:27:08 +0100

[diff] [blame]

59

RESIZE_NEAREST_NEIGHBOR (FLOAT32,QUANT8_ASYMM)

Francis Murtagh

074c25a

2019-07-22 16:40:57 +0100

[diff] [blame]

60

SOFTMAX (FLOAT32,QUANT8_ASYMM)

David Monahan

613b49c

2019-06-27 11:37:47 +0100

[diff] [blame]

61

TRANSPOSE_CONV_2D (FLOAT32,QUANT8_ASYMM)

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

62

63

--- Unsupported operators ---

64

Aron Virginas-Tar

2019-06-17 15:48:45 +0100

[diff] [blame]

65

The following AndroidNN HAL 1.0 operations are currently not supported:

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

66

67

DEPTH_TO_SPACE

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

68

EMBEDDING_LOOKUP

69

HASHTABLE_LOOKUP

70

LSH_PROJECTION

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

71

RNN

telsoa01

2018-03-09 13:51:08 +0000

[diff] [blame]

72

SVDF

73

Aron Virginas-Tar

2019-06-17 15:48:45 +0100

[diff] [blame]

74

The following AndroidNN HAL 1.2 operations are currently not supported:

75

76

CONCATENATION

77

LSTM

Aron Virginas-Tar