Skip to content

Instantly share code, notes, and snippets.

View gamble27's full-sized avatar
❄️
Knowledge is power

Olha Moskanova gamble27

❄️
Knowledge is power
6 followers · 13 following
View GitHub Profile
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_LReLU_128
Created July 24, 2019 13:17
128 LEAKYReLU output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
0%| | 0/1687 [00:00<?, ?it/s]-------------
Epoch 1:
99%|█████████▉| 1676/1687 [00:04<00:00, 432.05it/s]Train loss: 0.366999 | Train accuracy: 0.897044
100%|██████████| 1687/1687 [00:04<00:00, 356.18it/s]
100%|██████████| 187/187 [00:00<00:00, 1578.13it/s]
Val loss: 0.179367 | Val accuracy: 0.947861
-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_moment_128
Created July 24, 2019 13:14
128 MOMENTUM 0.9 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:06<00:00, 254.79it/s]
Train loss: 0.290504 | Train accuracy: 0.912882
81%|████████▏ | 152/187 [00:00<00:00, 1512.84it/s]Val loss: 0.132620 | Val accuracy: 0.961898
100%|██████████| 187/187 [00:00<00:00, 1517.45it/s]
0%| | 0/1687 [00:00<?, ?it/s]-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_lr_128
Created July 24, 2019 13:11
128 LEARNING RATE 0.01 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:04<00:00, 380.60it/s]
0%| | 0/187 [00:00<?, ?it/s]Train loss: 0.908129 | Train accuracy: 0.796792
100%|██████████| 187/187 [00:00<00:00, 1429.18it/s]
Val loss: 0.377350 | Val accuracy: 0.906250
0%| | 0/1687 [00:00<?, ?it/s]-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_batchsize_128
Created July 24, 2019 13:07
128 BATCHSIZE 64 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 843/843 [00:04<00:00, 205.36it/s]
0%| | 0/93 [00:00<?, ?it/s]Train loss: 0.463914 | Train accuracy: 0.875315
100%|██████████| 93/93 [00:00<00:00, 960.17it/s]
Val loss: 0.232220 | Val accuracy: 0.932460
0%| | 0/843 [00:00<?, ?it/s]-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_baseline_128
Created July 24, 2019 13:03
BASELINE 128 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:05<00:00, 304.09it/s]
Train loss: 0.366369 | Train accuracy: 0.897081
57%|█████▋ | 107/187 [00:00<00:00, 1066.80it/s]Val loss: 0.178036 | Val accuracy: 0.947861
100%|██████████| 187/187 [00:00<00:00, 1194.02it/s]
0%| | 0/1687 [00:00<?, ?it/s]-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_LReLU
Last active July 24, 2019 12:41
LeakyReLU output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:03<00:00, 551.87it/s]
0%| | 0/187 [00:00<?, ?it/s]Train loss: 0.444938 | Train accuracy: 0.871128
Val loss: 0.261497 | Val accuracy: 0.922627
100%|██████████| 187/187 [00:00<00:00, 2306.46it/s]
-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_moment
Created July 24, 2019 12:20
MOMENTUM 0.9 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:02<00:00, 586.44it/s]
0%| | 0/187 [00:00<?, ?it/s]Train loss: 0.616156 | Train accuracy: 0.821299
100%|██████████| 187/187 [00:00<00:00, 2232.56it/s]
Val loss: 0.469204 | Val accuracy: 0.870321
0%| | 0/1687 [00:00<?, ?it/s]-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_lr
Created July 24, 2019 12:17
LEARNING RATE 0.01 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:04<00:00, 339.18it/s]
0%| | 0/187 [00:00<?, ?it/s]Train loss: 1.060107 | Train accuracy: 0.720417
100%|██████████| 187/187 [00:00<00:00, 2490.51it/s]
Val loss: 0.422035 | Val accuracy: 0.895388
-------------
Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_batchsize
Created July 24, 2019 12:10
BATCHSIZE 64 output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 843/843 [00:09<00:00, 90.76it/s]
Train loss: 0.528406 | Train accuracy: 0.851016
100%|██████████| 93/93 [00:00<00:00, 1222.07it/s]
Val loss: 0.270610 | Val accuracy: 0.922715
-------------
0%| | 0/843 [00:00<?, ?it/s]Epoch 2:
@gamble27
gamble27 / MIT_MNIST_nnet_fc_outp_baseline
Last active July 24, 2019 12:11
BASELINE output of simple NN created for MNIST dataset by MITx course
/home/olga/Projects/ML_MIT/venv/bin/python /home/olga/Projects/ML_MIT/project2_mnist/part2_mnist/nnet_fc.py
-------------
Epoch 1:
100%|██████████| 1687/1687 [00:06<00:00, 262.83it/s]
Train loss: 0.445684 | Train accuracy: 0.870758
100%|██████████| 187/187 [00:00<00:00, 2348.46it/s]
Val loss: 0.260342 | Val accuracy: 0.923630
-------------
Epoch 2: