CNN14-cat

74.8% passed tests (460 passed / 155 failed).

Tests overview

Topic

Passed Tests

Correctness classification

38.0%

Correctness distribution

52.5%

Correctness speaker average

41.7%

Correctness speaker ranking

37.5%

Fairness accent

96.0%

Fairness language

100.0%

Fairness linguistic sentiment

97.9%

Fairness pitch

100.0%

Fairness sex

100.0%

Robustness background noise

56.7%

Robustness low quality phone

90.0%

Robustness recording condition

0.0%

Robustness simulated recording condition

0.0%

Robustness small changes

74.0%

Robustness spectral tilt

65.0%

Model parameters

Entry

Value

data

{‘train’: {‘msppodcast’: {‘version’: ‘2.6.0’, ‘table’: ‘emotion.categories.train.gold_standard’}}, ‘dev’: {‘msppodcast’: {‘version’: ‘2.6.0’, ‘table’: ‘emotion.categories.dev.gold_standard’}}}

model

cnn14

sampling_rate

16000

features

LogMel64

softmax

True

quantization

False

Model information

Entry

Value

Author

aderington

Date

2023-04-14

Name

onnx

Subgroup

ser.basic4.audpann.cnn14

Version

1.0.1

Min_length

500

Batch_size

16

Learning_rate

0.001

Best_epoch

34

Seed

3

Sample

{‘emodb’: {‘version’: ‘1.1.1’, ‘flavor’: {‘bit_depth’: None, ‘channels’: None, ‘format’: ‘wav’, ‘mixdown’: True, ‘sampling_rate’: 16000}, ‘data’: {‘wav/03a01Fa.wav’: [[0.8074951171875, 0.03554034233093262, 0.14863066375255585, 0.008333970792591572]], ‘wav/08a01Ab.wav’: [[0.5907111167907715, 0.15763787925243378, 0.15374456346035004, 0.0979064479470253]]}}}

Model_ids

{‘torch’: ‘97aa2de2-1.0.1’, ‘onnx_quant’: ‘1ce05bad-1.0.1’}