DBlindForDiscriminativeAI - Impulses

Experiments (4 / 10)

Name	Input	DSP Blocks	Learn Blocks	Tr_Proc	f32_V_Acc	f32_T_Acc	i8_V_Acc	i8_T_Acc	f32_Latency	f32_RAM	f32_Flash	i8_Latency	i8_RAM	i8_Flash
Impulse #1	1,000ms.	MFE , Spectrogram	MobileNetV2 0.35, Classifier	CPU	-	-	-	-	-	-	-	-	-	-	View impulse
Impulse #2	968ms.	MFCC , MFE , Spectrogram , Syntiant , Raw data	Classifier 1, Classifier 2, Classifier Ours, Classifier Spectogram, Classifier Syntiant, Classifier Raw Data	CPU	80.1%	-	80.1%	-	1799 ms.	61.8K	3.4M	629 ms.	60.5K	994.8K	View impulse
Impulse #3	1,000ms.	MFE , Spectrogram	Classifier	CPU	82.6%	-	83.8%	-	777 ms.	32.9K	37.5K	452 ms.	27.9K	34.2K	View impulse
Impulse #38	1,000ms.	MFE	Classifier	CPU	96.5%	-	96.5%	-	695 ms.	32.9K	31.9K	259 ms.	19.8K	31.8K	View impulse

Impulse information

Tags - Impulse tags

Source - Impulse source (Manual vs EON Tuner)

Created by - The user who originally created this impulse

Key metrics

Tr_Proc - Training processor

f32_V_Acc - Accuracy (float32, validation set)

f32_T_Acc - Accuracy (float32, test set)

i8_V_Acc - Accuracy (int8, validation set)

i8_T_Acc - Accuracy (int8, test set)

f32_Latency - float32 Latency for the complete impulse.

f32_RAM - float32 RAM for the complete impulse.

f32_Flash - float32 Flash for the complete impulse.

i8_Latency - int8 Latency for the complete impulse.

i8_RAM - int8 RAM for the complete impulse.

i8_Flash - int8 Flash for the complete impulse.

Input block config

Win_Size - Window size

Win_Incr - Window increase

Freq - Frequency

Zero_Pad - Zero pad data

Processing block config

P_1_Norm - Normalization

P_1_Fra_len - Mel-filterbank energy features > Frame length (Processing Block #1)

P_1_Fra_str - Mel-filterbank energy features > Frame stride (Processing Block #1)

P_1_Fil_num - Mel-filterbank energy features > Filter number (Processing Block #1)

P_1_FFT_len - Mel-filterbank energy features > FFT length (Processing Block #1)

P_1_Low_fre - Mel-filterbank energy features > Low frequency (Processing Block #1)

P_1_Noi_flo - Normalization > Noise floor (dB) (Processing Block #1)

P_1_Noc - Mel Frequency Cepstral Coefficients > Number of coefficients (Processing Block #1)

P_1_Nws - Mel Frequency Cepstral Coefficients > Normalization window size (Processing Block #1)

P_1_Coe - Pre-emphasis > Coefficient (Processing Block #1)

P_2_Norm - Normalization

P_2_Fra_len - Spectrogram > Frame length (Processing Block #2)

P_2_Fra_str - Spectrogram > Frame stride (Processing Block #2)

P_2_FFT_len - Spectrogram > FFT length (Processing Block #2)

P_2_Noi_flo - Normalization > Noise floor (dB) (Processing Block #2)

P_2_Fil_num - Mel-filterbank energy features > Filter number (Processing Block #2)

P_2_Low_fre - Mel-filterbank energy features > Low frequency (Processing Block #2)

P_3_Norm - Normalization

P_3_Fra_len - Spectrogram > Frame length (Processing Block #3)

P_3_Fra_str - Spectrogram > Frame stride (Processing Block #3)

P_3_FFT_len - Spectrogram > FFT length (Processing Block #3)

P_3_Noi_flo - Normalization > Noise floor (dB) (Processing Block #3)

P_4_Norm - Normalization

P_4_Fra_len - Log Mel filterbank energy features > Frame length (Processing Block #4)

P_4_Fra_str - Log Mel filterbank energy features > Frame stride (Processing Block #4)

P_4_Fil_num - Log Mel filterbank energy features > Filter number (Processing Block #4)

P_4_FFT_len - Log Mel filterbank energy features > FFT length (Processing Block #4)

P_4_Low_fre - Log Mel filterbank energy features > Low frequency (Processing Block #4)

P_4_Hig_fre - Log Mel filterbank energy features > High frequency (Processing Block #4)

P_4_Coe - Preemphasis > Coefficient (Processing Block #4)

P_4_Fea_ext - Chip > Features extractor (Processing Block #4)

P_5_Norm - Normalization

P_5_Sca_axe - Scaling > Scale axes (Processing Block #5)

Learn block config

L_1_Model - ML Model architecture (Learn Block #1)

L_1_Tc - Training cycles (Learn Block #1)

L_1_Lr - Learning rate (Learn Block #1)

L_1_Lo - Use learned optimizer? (Learn Block #1)

L_1_Do - Dropout (Learn Block #1)

L_1_Ne - Neurons (Learn Block #1)

L_1_Acw - Auto-weight classes (Learn Block #1)

L_1_Data_Aug - Data augmentation (Learn Block #1)

L_1_Val_Size - Validation set size (Learn Block #1)

L_1_Val_Key - Split train/validation set on metadata key (Learn Block #1)

L_1_Mode - Visual / expert mode (Learn Block #1)

L_1_Thr_Mcr - Minimum confidence rating (Learn Block #1)

L_2_Model - ML Model architecture (Learn Block #2)

L_2_Tc - Training cycles (Learn Block #2)

L_2_Lr - Learning rate (Learn Block #2)

L_2_Lo - Use learned optimizer? (Learn Block #2)

L_2_Acw - Auto-weight classes (Learn Block #2)

L_2_Data_Aug - Data augmentation (Learn Block #2)

L_2_Val_Size - Validation set size (Learn Block #2)

L_2_Val_Key - Split train/validation set on metadata key (Learn Block #2)

L_2_Mode - Visual / expert mode (Learn Block #2)

L_3_Val_Size - Validation set size (Learn Block #3)

L_3_Val_Key - Split train/validation set on metadata key (Learn Block #3)

L_3_Mode - Visual / expert mode (Learn Block #3)

L_4_Model - ML Model architecture (Learn Block #4)

L_4_Tc - Training cycles (Learn Block #4)

L_4_Lr - Learning rate (Learn Block #4)

L_4_Lo - Use learned optimizer? (Learn Block #4)

L_4_Acw - Auto-weight classes (Learn Block #4)

L_4_Data_Aug - Data augmentation (Learn Block #4)

L_4_Val_Size - Validation set size (Learn Block #4)

L_4_Val_Key - Split train/validation set on metadata key (Learn Block #4)

L_4_Mode - Visual / expert mode (Learn Block #4)

L_5_Model - ML Model architecture (Learn Block #5)

L_5_Tc - Training cycles (Learn Block #5)

L_5_Lr - Learning rate (Learn Block #5)

L_5_Lo - Use learned optimizer? (Learn Block #5)

L_5_Acw - Auto-weight classes (Learn Block #5)

L_5_Data_Aug - Data augmentation (Learn Block #5)

L_5_Val_Size - Validation set size (Learn Block #5)

L_5_Val_Key - Split train/validation set on metadata key (Learn Block #5)

L_5_Mode - Visual / expert mode (Learn Block #5)

L_6_Model - ML Model architecture (Learn Block #6)

L_6_Tc - Training cycles (Learn Block #6)

L_6_Lr - Learning rate (Learn Block #6)

L_6_Lo - Use learned optimizer? (Learn Block #6)

L_6_Acw - Auto-weight classes (Learn Block #6)

L_6_Data_Aug - Data augmentation (Learn Block #6)

L_6_Val_Size - Validation set size (Learn Block #6)

L_6_Val_Key - Split train/validation set on metadata key (Learn Block #6)

L_6_Mode - Visual / expert mode (Learn Block #6)

Accuracy metrics

f32_V_AUC - float32 Area under ROC Curve (validation set)

f32_V_WA_P - float32 Weighted average Precision (validation set)

f32_V_WA_R - float32 Weighted average Recall (validation set)

f32_V_WA_F1 - float32 Weighted average F1 score (validation set)

f32_T_AUC - float32 Area under ROC Curve (test set)

f32_T_WA_P - float32 Weighted average Precision (test set)

f32_T_WA_R - float32 Weighted average Recall (test set)

f32_T_WA_F1 - float32 Weighted average F1 score (test set)

i8_V_AUC - int8 Area under ROC Curve (validation set)

i8_V_WA_P - int8 Weighted average Precision (validation set)

i8_V_WA_R - int8 Weighted average Recall (validation set)

i8_V_WA_F1 - int8 Weighted average F1 score (validation set)

Post-processing block config

No columns

Target device

Define your target device requirements to inform model optimizations and performance calculations. No device yet? Use the default settings which you can change at any time.

Target device

Processor family

Clock rate

| MHz

Max

Custom device name

Application budget

Specify the available RAM and ROM for the model's operation, along with the maximum allowed latency for your specific application. Not sure yet? Start with the defaults and modify them later on.

RAM

| KB

Max

ROM

| KB

Max

Latency

| ms

Max

Experiments (4 / 10)

Impulse #1

Impulse #2

Impulse #3

Impulse #38