Counts the number of OPS in your Keras model.
Project description
KerOp: Keras-Opcounter
How to install
pip install kerop
-
Counts the number of OPS in your Keras model!
-
Visualizes the OPS at each of your layer, to find the bottleneck.
Supported layers: conv2d, fc, residual.
Usage
import kerop
# analyze FLOPS
layer_name, layer_flops, inshape, weights = kerop.profile(your_keras_model)
# visualize results
for name, flop, shape, weight in zip(layer_name, layer_flops, inshape, weights):
print("layer:", name, shape, " MegaFLOPS:", flop/1e6, " MegaWeights:", weight/1e6)
great thanks
@Lyken17
https://github.com/Lyken17/pytorch-OpCounter
Example: Count the ops in VGG16:
python example/vgg.py
Total params: 138,357,544
Trainable params: 138,357,544
Non-trainable params: 0
_________________________________________________________________
layer: block1_conv1 (None, 224, 224, 3) MegaFLOPS: 173.40832 MegaWeights: 0.001792
layer: block1_conv2 (None, 224, 224, 64) MegaFLOPS: 3699.376192 MegaWeights: 0.036928
layer: block2_conv1 (None, 112, 112, 64) MegaFLOPS: 1849.688192 MegaWeights: 0.073856
layer: block2_conv2 (None, 112, 112, 128) MegaFLOPS: 3699.376256 MegaWeights: 0.147584
layer: block3_conv1 (None, 56, 56, 128) MegaFLOPS: 1849.68832 MegaWeights: 0.295168
layer: block3_conv2 (None, 56, 56, 256) MegaFLOPS: 3699.376384 MegaWeights: 0.59008
layer: block3_conv3 (None, 56, 56, 256) MegaFLOPS: 3699.376384 MegaWeights: 0.59008
layer: block4_conv1 (None, 28, 28, 256) MegaFLOPS: 1849.688576 MegaWeights: 1.18016
layer: block4_conv2 (None, 28, 28, 512) MegaFLOPS: 3699.37664 MegaWeights: 2.359808
layer: block4_conv3 (None, 28, 28, 512) MegaFLOPS: 3699.37664 MegaWeights: 2.359808
layer: block5_conv1 (None, 14, 14, 512) MegaFLOPS: 924.844544 MegaWeights: 2.359808
layer: block5_conv2 (None, 14, 14, 512) MegaFLOPS: 924.844544 MegaWeights: 2.359808
layer: block5_conv3 (None, 14, 14, 512) MegaFLOPS: 924.844544 MegaWeights: 2.359808
layer: fc1 (None, 25088) MegaFLOPS: 205.524992 MegaWeights: 102.764544
layer: fc2 (None, 4096) MegaFLOPS: 33.558528 MegaWeights: 16.781312
Total FLOPS[GFLOPS]: 30.932349056
Count the ops in resnet50:
python example/resnet.py
Total params: 25,636,712
Trainable params: 25,583,592
Non-trainable params: 53,120
__________________________________________________________________________________________________
layer: conv1 (None, 230, 230, 3) MegaFLOPS: 497.683264 MegaWeights: 0.009472
layer: res2a_branch2a (None, 56, 56, 64) MegaFLOPS: 25.690176 MegaWeights: 0.00416
layer: res2a_branch2b (None, 56, 56, 64) MegaFLOPS: 231.211072 MegaWeights: 0.036928
layer: res2a_branch2c (None, 56, 56, 64) MegaFLOPS: 102.760704 MegaWeights: 0.01664
layer: res2a_branch1 (None, 56, 56, 64) MegaFLOPS: 102.760704 MegaWeights: 0.01664
layer: res2b_branch2a (None, 56, 56, 256) MegaFLOPS: 102.760512 MegaWeights: 0.016448
layer: res2b_branch2b (None, 56, 56, 64) MegaFLOPS: 231.211072 MegaWeights: 0.036928
layer: res2b_branch2c (None, 56, 56, 64) MegaFLOPS: 102.760704 MegaWeights: 0.01664
layer: res2c_branch2a (None, 56, 56, 256) MegaFLOPS: 102.760512 MegaWeights: 0.016448
layer: res2c_branch2b (None, 56, 56, 64) MegaFLOPS: 231.211072 MegaWeights: 0.036928
layer: res2c_branch2c (None, 56, 56, 64) MegaFLOPS: 102.760704 MegaWeights: 0.01664
layer: res3a_branch2a (None, 56, 56, 256) MegaFLOPS: 102.760576 MegaWeights: 0.032896
layer: res3a_branch2b (None, 28, 28, 128) MegaFLOPS: 231.211136 MegaWeights: 0.147584
layer: res3a_branch2c (None, 28, 28, 128) MegaFLOPS: 102.76096 MegaWeights: 0.066048
layer: res3a_branch1 (None, 56, 56, 256) MegaFLOPS: 411.042304 MegaWeights: 0.131584
layer: res3b_branch2a (None, 28, 28, 512) MegaFLOPS: 102.760576 MegaWeights: 0.065664
layer: res3b_branch2b (None, 28, 28, 128) MegaFLOPS: 231.211136 MegaWeights: 0.147584
layer: res3b_branch2c (None, 28, 28, 128) MegaFLOPS: 102.76096 MegaWeights: 0.066048
layer: res3c_branch2a (None, 28, 28, 512) MegaFLOPS: 102.760576 MegaWeights: 0.065664
layer: res3c_branch2b (None, 28, 28, 128) MegaFLOPS: 231.211136 MegaWeights: 0.147584
layer: res3c_branch2c (None, 28, 28, 128) MegaFLOPS: 102.76096 MegaWeights: 0.066048
layer: res3d_branch2a (None, 28, 28, 512) MegaFLOPS: 102.760576 MegaWeights: 0.065664
layer: res3d_branch2b (None, 28, 28, 128) MegaFLOPS: 231.211136 MegaWeights: 0.147584
layer: res3d_branch2c (None, 28, 28, 128) MegaFLOPS: 102.76096 MegaWeights: 0.066048
layer: res4a_branch2a (None, 28, 28, 512) MegaFLOPS: 102.760704 MegaWeights: 0.131328
layer: res4a_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4a_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res4a_branch1 (None, 28, 28, 512) MegaFLOPS: 411.042816 MegaWeights: 0.525312
layer: res4b_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.760704 MegaWeights: 0.2624
layer: res4b_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4b_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res4c_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.760704 MegaWeights: 0.2624
layer: res4c_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4c_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res4d_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.760704 MegaWeights: 0.2624
layer: res4d_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4d_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res4e_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.760704 MegaWeights: 0.2624
layer: res4e_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4e_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res4f_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.760704 MegaWeights: 0.2624
layer: res4f_branch2b (None, 14, 14, 256) MegaFLOPS: 231.211264 MegaWeights: 0.59008
layer: res4f_branch2c (None, 14, 14, 256) MegaFLOPS: 102.761472 MegaWeights: 0.263168
layer: res5a_branch2a (None, 14, 14, 1024) MegaFLOPS: 102.76096 MegaWeights: 0.5248
layer: res5a_branch2b (None, 7, 7, 512) MegaFLOPS: 231.21152 MegaWeights: 2.359808
layer: res5a_branch2c (None, 7, 7, 512) MegaFLOPS: 102.762496 MegaWeights: 1.050624
layer: res5a_branch1 (None, 14, 14, 1024) MegaFLOPS: 411.04384 MegaWeights: 2.0992
layer: res5b_branch2a (None, 7, 7, 2048) MegaFLOPS: 102.76096 MegaWeights: 1.049088
layer: res5b_branch2b (None, 7, 7, 512) MegaFLOPS: 231.21152 MegaWeights: 2.359808
layer: res5b_branch2c (None, 7, 7, 512) MegaFLOPS: 102.762496 MegaWeights: 1.050624
layer: res5c_branch2a (None, 7, 7, 2048) MegaFLOPS: 102.76096 MegaWeights: 1.049088
layer: res5c_branch2b (None, 7, 7, 512) MegaFLOPS: 231.21152 MegaWeights: 2.359808
layer: res5c_branch2c (None, 7, 7, 512) MegaFLOPS: 102.762496 MegaWeights: 1.050624
layer: fc1000 (None, 2048) MegaFLOPS: 4.097 MegaWeights: 2.049
Total FLOPS[GFLOPS]: 8.748332712
densenet
python example/densenet.py
layer: conv1/conv (None, 230, 230, 3) MegaFLOPS: 236.027904 MegaWeights: 0.009408
layer: conv2_block1_1_conv (None, 56, 56, 64) MegaFLOPS: 51.380224 MegaWeights: 0.008192
layer: conv2_block1_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: conv2_block2_1_conv (None, 56, 56, 96) MegaFLOPS: 77.070336 MegaWeights: 0.012288
layer: conv2_block2_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: conv2_block3_1_conv (None, 56, 56, 128) MegaFLOPS: 102.760448 MegaWeights: 0.016384
layer: conv2_block3_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: conv2_block4_1_conv (None, 56, 56, 160) MegaFLOPS: 128.45056 MegaWeights: 0.02048
layer: conv2_block4_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: conv2_block5_1_conv (None, 56, 56, 192) MegaFLOPS: 154.140672 MegaWeights: 0.024576
layer: conv2_block5_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: conv2_block6_1_conv (None, 56, 56, 224) MegaFLOPS: 179.830784 MegaWeights: 0.028672
layer: conv2_block6_2_conv (None, 56, 56, 128) MegaFLOPS: 231.211008 MegaWeights: 0.036864
layer: pool2_conv (None, 56, 56, 256) MegaFLOPS: 205.520896 MegaWeights: 0.032768
layer: conv3_block1_1_conv (None, 28, 28, 128) MegaFLOPS: 25.690112 MegaWeights: 0.016384
layer: conv3_block1_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block2_1_conv (None, 28, 28, 160) MegaFLOPS: 32.11264 MegaWeights: 0.02048
layer: conv3_block2_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block3_1_conv (None, 28, 28, 192) MegaFLOPS: 38.535168 MegaWeights: 0.024576
layer: conv3_block3_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block4_1_conv (None, 28, 28, 224) MegaFLOPS: 44.957696 MegaWeights: 0.028672
layer: conv3_block4_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block5_1_conv (None, 28, 28, 256) MegaFLOPS: 51.380224 MegaWeights: 0.032768
layer: conv3_block5_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block6_1_conv (None, 28, 28, 288) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block6_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block7_1_conv (None, 28, 28, 320) MegaFLOPS: 64.22528 MegaWeights: 0.04096
layer: conv3_block7_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block8_1_conv (None, 28, 28, 352) MegaFLOPS: 70.647808 MegaWeights: 0.045056
layer: conv3_block8_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block9_1_conv (None, 28, 28, 384) MegaFLOPS: 77.070336 MegaWeights: 0.049152
layer: conv3_block9_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block10_1_conv (None, 28, 28, 416) MegaFLOPS: 83.492864 MegaWeights: 0.053248
layer: conv3_block10_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block11_1_conv (None, 28, 28, 448) MegaFLOPS: 89.915392 MegaWeights: 0.057344
layer: conv3_block11_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: conv3_block12_1_conv (None, 28, 28, 480) MegaFLOPS: 96.33792 MegaWeights: 0.06144
layer: conv3_block12_2_conv (None, 28, 28, 128) MegaFLOPS: 57.802752 MegaWeights: 0.036864
layer: pool3_conv (None, 28, 28, 512) MegaFLOPS: 205.520896 MegaWeights: 0.131072
layer: conv4_block1_1_conv (None, 14, 14, 256) MegaFLOPS: 12.845056 MegaWeights: 0.032768
layer: conv4_block1_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block2_1_conv (None, 14, 14, 288) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block2_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block3_1_conv (None, 14, 14, 320) MegaFLOPS: 16.05632 MegaWeights: 0.04096
layer: conv4_block3_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block4_1_conv (None, 14, 14, 352) MegaFLOPS: 17.661952 MegaWeights: 0.045056
layer: conv4_block4_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block5_1_conv (None, 14, 14, 384) MegaFLOPS: 19.267584 MegaWeights: 0.049152
layer: conv4_block5_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block6_1_conv (None, 14, 14, 416) MegaFLOPS: 20.873216 MegaWeights: 0.053248
layer: conv4_block6_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block7_1_conv (None, 14, 14, 448) MegaFLOPS: 22.478848 MegaWeights: 0.057344
layer: conv4_block7_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block8_1_conv (None, 14, 14, 480) MegaFLOPS: 24.08448 MegaWeights: 0.06144
layer: conv4_block8_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block9_1_conv (None, 14, 14, 512) MegaFLOPS: 25.690112 MegaWeights: 0.065536
layer: conv4_block9_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block10_1_conv (None, 14, 14, 544) MegaFLOPS: 27.295744 MegaWeights: 0.069632
layer: conv4_block10_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block11_1_conv (None, 14, 14, 576) MegaFLOPS: 28.901376 MegaWeights: 0.073728
layer: conv4_block11_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block12_1_conv (None, 14, 14, 608) MegaFLOPS: 30.507008 MegaWeights: 0.077824
layer: conv4_block12_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block13_1_conv (None, 14, 14, 640) MegaFLOPS: 32.11264 MegaWeights: 0.08192
layer: conv4_block13_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block14_1_conv (None, 14, 14, 672) MegaFLOPS: 33.718272 MegaWeights: 0.086016
layer: conv4_block14_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block15_1_conv (None, 14, 14, 704) MegaFLOPS: 35.323904 MegaWeights: 0.090112
layer: conv4_block15_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block16_1_conv (None, 14, 14, 736) MegaFLOPS: 36.929536 MegaWeights: 0.094208
layer: conv4_block16_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block17_1_conv (None, 14, 14, 768) MegaFLOPS: 38.535168 MegaWeights: 0.098304
layer: conv4_block17_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block18_1_conv (None, 14, 14, 800) MegaFLOPS: 40.1408 MegaWeights: 0.1024
layer: conv4_block18_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block19_1_conv (None, 14, 14, 832) MegaFLOPS: 41.746432 MegaWeights: 0.106496
layer: conv4_block19_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block20_1_conv (None, 14, 14, 864) MegaFLOPS: 43.352064 MegaWeights: 0.110592
layer: conv4_block20_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block21_1_conv (None, 14, 14, 896) MegaFLOPS: 44.957696 MegaWeights: 0.114688
layer: conv4_block21_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block22_1_conv (None, 14, 14, 928) MegaFLOPS: 46.563328 MegaWeights: 0.118784
layer: conv4_block22_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block23_1_conv (None, 14, 14, 960) MegaFLOPS: 48.16896 MegaWeights: 0.12288
layer: conv4_block23_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block24_1_conv (None, 14, 14, 992) MegaFLOPS: 49.774592 MegaWeights: 0.126976
layer: conv4_block24_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block25_1_conv (None, 14, 14, 1024) MegaFLOPS: 51.380224 MegaWeights: 0.131072
layer: conv4_block25_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block26_1_conv (None, 14, 14, 1056) MegaFLOPS: 52.985856 MegaWeights: 0.135168
layer: conv4_block26_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block27_1_conv (None, 14, 14, 1088) MegaFLOPS: 54.591488 MegaWeights: 0.139264
layer: conv4_block27_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block28_1_conv (None, 14, 14, 1120) MegaFLOPS: 56.19712 MegaWeights: 0.14336
layer: conv4_block28_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block29_1_conv (None, 14, 14, 1152) MegaFLOPS: 57.802752 MegaWeights: 0.147456
layer: conv4_block29_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block30_1_conv (None, 14, 14, 1184) MegaFLOPS: 59.408384 MegaWeights: 0.151552
layer: conv4_block30_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block31_1_conv (None, 14, 14, 1216) MegaFLOPS: 61.014016 MegaWeights: 0.155648
layer: conv4_block31_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: conv4_block32_1_conv (None, 14, 14, 1248) MegaFLOPS: 62.619648 MegaWeights: 0.159744
layer: conv4_block32_2_conv (None, 14, 14, 128) MegaFLOPS: 14.450688 MegaWeights: 0.036864
layer: pool4_conv (None, 14, 14, 1280) MegaFLOPS: 321.1264 MegaWeights: 0.8192
layer: conv5_block1_1_conv (None, 7, 7, 640) MegaFLOPS: 8.02816 MegaWeights: 0.08192
layer: conv5_block1_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block2_1_conv (None, 7, 7, 672) MegaFLOPS: 8.429568 MegaWeights: 0.086016
layer: conv5_block2_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block3_1_conv (None, 7, 7, 704) MegaFLOPS: 8.830976 MegaWeights: 0.090112
layer: conv5_block3_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block4_1_conv (None, 7, 7, 736) MegaFLOPS: 9.232384 MegaWeights: 0.094208
layer: conv5_block4_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block5_1_conv (None, 7, 7, 768) MegaFLOPS: 9.633792 MegaWeights: 0.098304
layer: conv5_block5_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block6_1_conv (None, 7, 7, 800) MegaFLOPS: 10.0352 MegaWeights: 0.1024
layer: conv5_block6_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block7_1_conv (None, 7, 7, 832) MegaFLOPS: 10.436608 MegaWeights: 0.106496
layer: conv5_block7_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block8_1_conv (None, 7, 7, 864) MegaFLOPS: 10.838016 MegaWeights: 0.110592
layer: conv5_block8_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block9_1_conv (None, 7, 7, 896) MegaFLOPS: 11.239424 MegaWeights: 0.114688
layer: conv5_block9_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block10_1_conv (None, 7, 7, 928) MegaFLOPS: 11.640832 MegaWeights: 0.118784
layer: conv5_block10_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block11_1_conv (None, 7, 7, 960) MegaFLOPS: 12.04224 MegaWeights: 0.12288
layer: conv5_block11_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block12_1_conv (None, 7, 7, 992) MegaFLOPS: 12.443648 MegaWeights: 0.126976
layer: conv5_block12_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block13_1_conv (None, 7, 7, 1024) MegaFLOPS: 12.845056 MegaWeights: 0.131072
layer: conv5_block13_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block14_1_conv (None, 7, 7, 1056) MegaFLOPS: 13.246464 MegaWeights: 0.135168
layer: conv5_block14_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block15_1_conv (None, 7, 7, 1088) MegaFLOPS: 13.647872 MegaWeights: 0.139264
layer: conv5_block15_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block16_1_conv (None, 7, 7, 1120) MegaFLOPS: 14.04928 MegaWeights: 0.14336
layer: conv5_block16_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block17_1_conv (None, 7, 7, 1152) MegaFLOPS: 14.450688 MegaWeights: 0.147456
layer: conv5_block17_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block18_1_conv (None, 7, 7, 1184) MegaFLOPS: 14.852096 MegaWeights: 0.151552
layer: conv5_block18_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block19_1_conv (None, 7, 7, 1216) MegaFLOPS: 15.253504 MegaWeights: 0.155648
layer: conv5_block19_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block20_1_conv (None, 7, 7, 1248) MegaFLOPS: 15.654912 MegaWeights: 0.159744
layer: conv5_block20_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block21_1_conv (None, 7, 7, 1280) MegaFLOPS: 16.05632 MegaWeights: 0.16384
layer: conv5_block21_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block22_1_conv (None, 7, 7, 1312) MegaFLOPS: 16.457728 MegaWeights: 0.167936
layer: conv5_block22_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block23_1_conv (None, 7, 7, 1344) MegaFLOPS: 16.859136 MegaWeights: 0.172032
layer: conv5_block23_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block24_1_conv (None, 7, 7, 1376) MegaFLOPS: 17.260544 MegaWeights: 0.176128
layer: conv5_block24_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block25_1_conv (None, 7, 7, 1408) MegaFLOPS: 17.661952 MegaWeights: 0.180224
layer: conv5_block25_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block26_1_conv (None, 7, 7, 1440) MegaFLOPS: 18.06336 MegaWeights: 0.18432
layer: conv5_block26_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block27_1_conv (None, 7, 7, 1472) MegaFLOPS: 18.464768 MegaWeights: 0.188416
layer: conv5_block27_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block28_1_conv (None, 7, 7, 1504) MegaFLOPS: 18.866176 MegaWeights: 0.192512
layer: conv5_block28_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block29_1_conv (None, 7, 7, 1536) MegaFLOPS: 19.267584 MegaWeights: 0.196608
layer: conv5_block29_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block30_1_conv (None, 7, 7, 1568) MegaFLOPS: 19.668992 MegaWeights: 0.200704
layer: conv5_block30_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block31_1_conv (None, 7, 7, 1600) MegaFLOPS: 20.0704 MegaWeights: 0.2048
layer: conv5_block31_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: conv5_block32_1_conv (None, 7, 7, 1632) MegaFLOPS: 20.471808 MegaWeights: 0.208896
layer: conv5_block32_2_conv (None, 7, 7, 128) MegaFLOPS: 3.612672 MegaWeights: 0.036864
layer: fc1000 (None, 1664) MegaFLOPS: 3.329 MegaWeights: 1.665
Total FLOPS[GFLOPS]: 6.719687656
still under development! :)
Numbers are still buggy, plz wait till I fix this.
Right now, supports conv2d and dense only.
Doesn't fully count the activations yet.
We count 1 MAC as 2 FLOPS and 1 ADD as 1 FLOPS.
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
File details
Details for the file kerop-0.1.4-2010260808.tar.gz.
File metadata
- Download URL: kerop-0.1.4-2010260808.tar.gz
- Upload date:
- Size: 6.5 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.2.0 pkginfo/1.5.0.1 requests/2.22.0 setuptools/45.2.0.post20200210 requests-toolbelt/0.9.1 tqdm/4.42.1 CPython/3.6.10
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
94671484ad66f74eaf2d4c6be6b5eb4811a87b18a4cddd0f31b786c5824cc299
|
|
| MD5 |
5de6d9bc1012cf99460140669d446d5e
|
|
| BLAKE2b-256 |
150ee8c06d41f2a584890b6631d3cf32a20cbfc1833c4d66424525bccd52bffc
|
