waleedka / hiddenlayer Goto Github PK
View Code? Open in Web Editor NEWNeural network graphs and training metrics for PyTorch, Tensorflow, and Keras.
License: MIT License
hiddenlayer's People
Contributors
Stargazers
Watchers
Forkers
ml-lab vvkishere tarrysingh ghzhangnj skyjiao mengrang albert0147 walden2013 lixiangideal zhengxiawu jiapei100 ringwraith scbwin ss18 uptodiff chineral aaronhd soonhwan-kwon alixing aust-hansen liyaangy oppa3109 yoyokitartora amseej ruhollah2 euwen chancing0 fulin-wei undercontroller beshining o7s8r6 savourylie zengxiaoqing jiaobingle wang1104014663 zaytiamo zhipengliu6 gzoumpourlis wiwi emreozkose michael-hsu pyadmell barbecacov tongluoiupui truong225 adrianbzg nanaj1 antyan001 yanghaocsg pandinosaurus kinlance thomaslin1990 ammieqi paul028 creatist thabet1983 phamcuong92 christinaliang martslaaf cxz chsafouane mlearnx qzcmask alzayats stargazeryuan zhenguoyuan wangzheallen sweden1003 chlapec flowtcw wenhengqiu fengjunxi lovepug-xc quxiaofeng changxiangshi muxinghan rizalespe mgc26 manik-hossain sanyalington gaogg cosgriffc wsc12358 mocuto byfaith nguyenducnhaty shayan-taheri ifighting xiaozhuka jimmy-ksu janysunny jschubnell praveen94 chy7074646 wut0n9 tectal haimiaom lehaifeng imjerrychen sskramhiddenlayer's Issues
Drawing error
Why did I use your code to draw the picture in a mess? I don't know where it is wrong because I just copy your code on github.
Save network plot to file
Hello,
I saw there is a possibility to save training stat plots into png files, but I couldn't find an option to save the actual NN architecture plot into a file, is this currently possible?
Thanks
UnboundLocalError: local variable 'FRAMEWORK_TRANSFORMS' referenced before assignment
when building a graph with "hl.build_graph(model)" ,model is a tensorflow net, it occurs:
hiddenlayer-master/hiddenlayer/graph.py in build_graph(model, args, input_names, transforms, framework_transforms)
147 if framework_transforms:
148 if framework_transforms == "default":
--> 149 framework_transforms = FRAMEWORK_TRANSFORMS
150 for t in framework_transforms:
151 g = t.apply(g)
UnboundLocalError: local variable 'FRAMEWORK_TRANSFORMS' referenced before assignment
Is it possible to control which part of the model graph is print
The graph is useful for visualizing model, but when the model is big, printing the entire model is not too helpful. Say I already have a structure of the model(a OrderDict or groups of layer), is that possible to print only part of a model graph.
I am using PyTorch with a pre-trained Resnet, I am only interested in the layer after the Resnet Encoder part. Ideally, I would like to print something like ResNet(Maybe just a manual input argument) + details of the interested layer groups.
ValueError on vgg16_bn unless I first train for one epoch
I'm able to build and display a graph right away on the torchvision alexnet model. But on the torchvision vgg16_bn model, if try to graph the model before training I get the error, ValueError: Expected more than 1 value per channel when training, got input size torch.Size([1, 1024]) Even though I passed in hl.build_graph(model, torch.zeros((1,3,224,224).cuda()). If I train the model for 1 epoch, then I'm able to build and display the graph properly.
Edit: Actually the torchvision model copied exactly works, but if I create a fastai learner with this model (which removes the head and adds a custom head with adaptive pooling) then I get this issue.
Edit 2: if I set the model to eval mode before trying to graph it works
Any idea why this is the case?
Is there a plan to support tf 2.0?
Hi, thanks for your work. Would you have a plan to support tf 2.0?
Thanks.
Keep `___init__.py` empty? (Codebase merge issue 1 of 3)
Argments Problem....
If my forward function has dict argment....what can I do?
For example....
import torch
import torch.nn as nn
import hiddenlayer as hl
import torch.jit as jit
class TestNet(nn.Module):
def __init__(self):
super(TestNet, self).__init__()
self.conv = nn.Conv2d(3, 3, 1, 1)
self.relu = nn.ReLU(True)
def forward(self, x , y, loss):
if loss['flag']:
x = self.conv(x)
else:
assert False
return x
x = torch.randn(1,3,224,224)
net = TestNet()
graph = hl.build_graph(net, (x, x, {'flag': True}))
And it will got wrong info as below...............
Traceback (most recent call last):
File "/home/gaozhihua/program/mmdetection/ignore_dir/1.py", line 27, in <module>
graph = hl.build_graph(net, (x, x, {'flag': True}))
File "/home/gaozhihua/program/hiddenlayer/hiddenlayer/graph.py", line 143, in build_graph
import_graph(g, model, args)
File "/home/gaozhihua/program/hiddenlayer/hiddenlayer/pytorch_builder.py", line 70, in import_graph
trace, out = torch.jit.get_trace_graph(model, args)
File "/home/gaozhihua/anaconda2/envs/open-mmlab/lib/python3.6/site-packages/torch/jit/__init__.py", line 196, in get_trace_graph
return LegacyTracedModule(f, _force_outplace)(*args, **kwargs)
File "/home/gaozhihua/anaconda2/envs/open-mmlab/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in __call__
result = self.forward(*input, **kwargs)
File "/home/gaozhihua/anaconda2/envs/open-mmlab/lib/python3.6/site-packages/torch/jit/__init__.py", line 242, in forward
in_vars, in_desc = _flatten(args)
RuntimeError: Only tuples, lists and Variables supported as JIT inputs, but got dict
What shuould I do?
@waleedka @FerumFlex @ss18
NotImplementedError
Getting NotImplementedError when trying to visualize graph for 3dcnn network in pytorch
Is latest `draw_graph()` function work in progress? (Codebase merge issue 3 of 3)
Shape visualization problem
Hi:
Thank you for you amazing work! But I have a problem about the visualization:How can I get more shape info? Can you tell me the solution?
best,
yuxiang
unable to handle batchnorm1d in pytorch
A simple example, no issues
When added batchnorm1d between layers,
Expected more than 1 value per channel when training, got input size torch.Size([1, 100])
It seems like it's treating batchnorm1d as batchnorm2d?
ValueError: model input param must be a PyTorch, TensorFlow, or Keras-with-TensorFlow-backend model
**I have the problem. I want to visualize this code, but my code had error:
ValueError: model input param must be a PyTorch, TensorFlow, or Keras-with-TensorFlow-backend model. I don't know, how do I set the input number? ( I am using the pytorch tool)
I did:
-
[ ]
" import hiddenlayer as hl input = torch.zeros([1, 3, 64, 64, 64]) model = VoxResNet() hl.build_graph(model, (input))"
I am following this code from GitHub: https://github.com/Ryo-Ito/brain_segmentation. Please, help me, visualizer. Thank you.**
import chainer
import chainer.functions as F
import chainer.links as L
import torch
class VoxResModule(chainer.Chain):
"""
Voxel Residual Module
input
BatchNormalization, ReLU
Conv 64, 3x3x3
BatchNormalization, ReLU
Conv 64, 3x3x3
output
"""
def init(self):
initW = chainer.initializers.HeNormal(scale=0.01)
super().init()
with self.init_scope():
self.bnorm1 = L.BatchNormalization(size=64)
self.conv1 = L.ConvolutionND(3, 64, 64, 3, pad=1, initialW=initW)
self.bnorm2 = L.BatchNormalization(size=64)
self.conv2 = L.ConvolutionND(3, 64, 64, 3, pad=1, initialW=initW)
def call(self, x):
h = F.relu(self.bnorm1(x))
h = self.conv1(h)
h = F.relu(self.bnorm2(h))
h = self.conv2(h)
return h + x
class VoxResNet(chainer.Chain):
"""Voxel Residual Network"""
def init(self, in_channels=1, n_classes=4):
init = chainer.initializers.HeNormal(scale=0.01)
super().init()
with self.init_scope():
self.conv1a = L.ConvolutionND(
3, in_channels, 32, 3, pad=1, initialW=init)
self.bnorm1a = L.BatchNormalization(32)
self.conv1b = L.ConvolutionND(
3, 32, 32, 3, pad=1, initialW=init)
self.bnorm1b = L.BatchNormalization(32)
self.conv1c = L.ConvolutionND(
3, 32, 64, 3, stride=2, pad=1, initialW=init)
self.voxres2 = VoxResModule()
self.voxres3 = VoxResModule()
self.bnorm3 = L.BatchNormalization(64)
self.conv4 = L.ConvolutionND(
3, 64, 64, 3, stride=2, pad=1, initialW=init)
self.voxres5 = VoxResModule()
self.voxres6 = VoxResModule()
self.bnorm6 = L.BatchNormalization(64)
self.conv7 = L.ConvolutionND(
3, 64, 64, 3, stride=2, pad=1, initialW=init)
self.voxres8 = VoxResModule()
self.voxres9 = VoxResModule()
self.c1deconv = L.DeconvolutionND(
3, 32, 32, 3, pad=1, initialW=init)
self.c1conv = L.ConvolutionND(
3, 32, n_classes, 3, pad=1, initialW=init)
self.c2deconv = L.DeconvolutionND(
3, 64, 64, 4, stride=2, pad=1, initialW=init)
self.c2conv = L.ConvolutionND(
3, 64, n_classes, 3, pad=1, initialW=init)
self.c3deconv = L.DeconvolutionND(
3, 64, 64, 6, stride=4, pad=1, initialW=init)
self.c3conv = L.ConvolutionND(
3, 64, n_classes, 3, pad=1, initialW=init)
self.c4deconv = L.DeconvolutionND(
3, 64, 64, 10, stride=8, pad=1, initialW=init)
self.c4conv = L.ConvolutionND(
3, 64, n_classes, 3, pad=1, initialW=init)
def call(self, x, train=False):
print(x.shape, '-------begin------------')
"""
calculate output of VoxResNet given input x
Parameters
----------
x : (batch_size, in_channels, xlen, ylen, zlen) ndarray
image to perform semantic segmentation
Returns
-------
proba: (batch_size, n_classes, xlen, ylen, zlen) ndarray
probability of each voxel belonging each class
elif train=True, returns list of logits
"""
with chainer.using_config("train", train):
h = self.conv1a(x)
h = F.relu(self.bnorm1a(h))
h = self.conv1b(h)
c1 = F.clipped_relu(self.c1deconv(h))
c1 = self.c1conv(c1)
h = F.relu(self.bnorm1b(h))
h = self.conv1c(h)
h = self.voxres2(h)
h = self.voxres3(h)
c2 = F.clipped_relu(self.c2deconv(h))
c2 = self.c2conv(c2)
h = F.relu(self.bnorm3(h))
h = self.conv4(h)
h = self.voxres5(h)
h = self.voxres6(h)
c3 = F.clipped_relu(self.c3deconv(h))
c3 = self.c3conv(c3)
h = F.relu(self.bnorm6(h))
h = self.conv7(h)
h = self.voxres8(h)
h = self.voxres9(h)
c4 = F.clipped_relu(self.c4deconv(h))
c4 = self.c4conv(c4)
c = c1 + c2 + c3 + c4
if train:
return [c1, c2, c3, c4, c]
else:
return F.softmax(c)
import hiddenlayer as hl
input = torch.zeros([1, 3, 64, 64, 64])
model = VoxResNet()
hl.build_graph(model, (input))
Failed to export an ONNX attribute, since it's not constant, please try to make things (e.g., kernel size) static if possible
Hi, I meet the problem using torchvision
model = torchvision.models.segmentation.deeplabv3_resnet101(pretrained=True)
model.eval()
trace, _ = torch.jit.get_trace_graph(model, torch.randn(1, 3, 224, 224))
torch.onnx._optimize_trace(trace, torch.onnx.OperatorExportTypes.ONNX)
Could you please give some hits to save this problem?
RuntimError: only tuples, list and Variables supported as JIT inputs, but got NoneType
`No handles with labels found to put in legend.` matplotlib messages
@waleedka , I haven't been able to successfully render ww plots with the TF watcher notebook.
If you look at cell 20, you'll see a matplotlib message that I get over and over. The web browser display flickers continuously and I don't get to ever see actual data in the plots.
Would you have time to investigate to figure out what's going on? Thx!
Horizontal view ?
Is it possible to get a horizontal view of the network graph ?
Assertation error running example pytorch notebooks
Hi,
I get the following error when I try and run your example pytorch notebooks:
~/code/development_tools/anaconda3/envs/fastai/lib/python3.7/site-packages/torch/onnx/symbolic.py in adaptive_avg_pool2d(g, input, output_size)
598 @parse_args('v', 'is')
599 def adaptive_avg_pool2d(g, input, output_size):
--> 600 assert output_size == [1, 1], "Only output_size=[1, 1] is supported"
601 return g.op("GlobalAveragePool", input)
602
AssertionError: Only output_size=[1, 1] is supported
I haven't modified anything, and I wonder if you have any idea what the source of this error is?
Regards,
Alex
hiddenlayer cannot identify that my module is indeed a torch.nn.Module
I have a personalised model class called let's say MyNet(Net), and it inherits from Net(nn.Module).
When I call hl.build_graph(model, ...), hiddenlayer then raises the exception:
- ValueError:
modelinput param must be a PyTorch, TensorFlow, or Keras-with-TensorFlow-backend model.
When I put everything inside only one class it works...
Possible to use with pytorch model generated by fastai?
Is it possible to use this on the models generated by fastai? If I run it on the pytorch model stored by the learner object, I get an error that it isn't a valid model.
ValueError: `model` input param must be a PyTorch, TensorFlow, or Keras-with-TensorFlow-backend model.
Problem with output size using FoldDuplicates.
Hello @waleedka !
I have just noticed, that when I use FoldDuplicates, I get wrong output size (On the graph the output size after the first module is shown, instead of the output size of the last module of the folded duplicates.)
Error while building graph for 3D ResNet
I am getting this error while building graph for 3D ResNet:
hl.build_graph(model,torch.zeros([1,3,16,112,12]))
RuntimeError: invalid argument 2: input image (T: 1 H: 4 W: 1) smaller than kernel size (kT: 1 kH: 4 kW: 4) at /opt/conda/conda-bld/pytorch_1549635019666/work/aten/src/THCUNN/generic/VolumetricAveragePooling.cu:57
When I want to visualize the efficientNet, a runtime error occurs.
import json
from PIL import Image
import torch
from torchvision import transforms
from torchviz import make_dot
import hiddenlayer as hl
from efficientnet_pytorch import EfficientNet
model = EfficientNet.from_pretrained('efficientnet-b3')
model.eval()
hl.build_graph(model, torch.zeros([1, 3, 224, 224]))
hl_graph
Traceback (most recent call last):
File "example.py", line 30, in
hl.build_graph(model, torch.zeros([1, 3, 224, 224]))
File "/home/tianhui/hiddenlayer/hiddenlayer/graph.py", line 143, in build_graph
import_graph(g, model, args)
File "/home/tianhui/hiddenlayer/hiddenlayer/pytorch_builder.py", line 71, in import_graph
torch.onnx._optimize_trace(trace, torch.onnx.OperatorExportTypes.ONNX)
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/init.py", line 40, in _optimize_trace
trace.set_graph(utils._optimize_graph(trace.graph(), operator_export_type))
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/utils.py", line 188, in _optimize_graph
graph = torch._C._jit_pass_onnx(graph, operator_export_type)
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/init.py", line 50, in _run_symbolic_function
return utils._run_symbolic_function(*args, **kwargs)
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/utils.py", line 589, in _run_symbolic_function
return fn(g, *inputs, **attrs)
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 130, in wrapper
args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 130, in
args = [_parse_arg(arg, arg_desc) for arg, arg_desc in zip(args, arg_descriptors)]
File "/home/lanqiang/anaconda3/lib/python3.6/site-packages/torch/onnx/symbolic.py", line 90, in _parse_arg
raise RuntimeError("Failed to export an ONNX attribute, "
RuntimeError: Failed to export an ONNX attribute, since it's not constant, please try to make things (e.g., kernel size) static if possible
TypeError: zeros(): argument 'out' (position 2) must be Tensor, not list
I want to visualizer a model with 3-input and have problem with feed 3-input. This model have three input, this filgure:
I conferenced this ideal of Waleed (https://github.com/waleedka)and I visualier 3 patch with 3 input [ 32,6, 25, 25] [ 32,6, 51, 51],[32, 6, 75, 75] with code line:
hl.build_graph(model, torch.zeros([32,6, 25, 25], [32,6, 51, 51],[32, 6, 75, 75]))
- But my code had the error: TypeError: zeros(): argument 'out' (position 2) must be Tensor, not list.
How do I fix this problem? (I also tried many ways, example:
#hl.build_graph(net, torch.zeros( 32,6, 25, 25), torch.zeros(32, 6, 51, 51), torch.zeros(32, 6, 75, 75))
#hl.build_graph(net, torch.zeros(( 32,6, 25, 25), ( 32,6, 51, 51), ( 32,6, 75, 75)))
#hl.build_graph(net, torch.zeros([( 32,6, 25, 25)],[( 32,6, 51, 51)],[( 32,6, 75, 75)])
)
In addition, I successed with the first path with code line:
hl.build_graph(model, torch.zeros([32,6, 25, 25]). And had figure:
.....................
2) And question more about window size of model visualize. Can we show full model visualize one window or save it ? as we see these below picture, I must to stride many time in oder to take a my model.
Thank you.
fail to run pytest
After building hiddenlayer, pytest -v failed at:
tests/test_pytorch_graph.py::TestPytorchGraph::test_graph [1] 29548 segmentation fault (core dumped) pytest -v
Any idea? Or should I ignore it anyway?
import hiddenlayer Segmentation fault (core dumped)
I tried to install hiddenlayer using pip and install it from source code, respectively. But when i import hiddenlayer, i met an error, wich is Segmentation fault (core dumped).
Do you have any idea about this bug?
What is number input for visulizer model?
I run this code in Jupyter Notebook，but one error occurs:
`import torch
import torchvision.models
import hidden layer as hl
model = torchvision.models.vgg16()
hl.build_graph(model, torch.zeros([1, 3, 224, 224]))'
And now, I want to feed three inputs are [ 32,6, 25, 25] [ 32,6, 51, 51],[32, 6, 75, 75]. I wrote : hl.build_graph(net, torch.zeros(
[ 32,6, 25, 25],
[ 32,6, 51, 51],
[32, 6, 75, 75]
)
But my code had the error: TypeError: zeros(): argument 'out' (position 2) must be Tensor, not list.
How do I fix this problem? (I also tried many ways, ex:
#hl.build_graph(net, torch.zeros( 32,6, 25, 25), torch.zeros(32, 6, 51, 51), torch.zeros(32, 6, 75, 75))
#hl.build_graph(net, torch.zeros(( 32,6, 25, 25), ( 32,6, 51, 51), ( 32,6, 75, 75)))
#hl.build_graph(net, torch.zeros([( 32,6, 25, 25)],[( 32,6, 51, 51)],[( 32,6, 75, 75)])
)
Thank you.
module 'torch.onnx' has no attribute 'OperatorExportTypes'
I run this code in Jupyter Notebook，but one error occurs:
`
import torch
import torchvision.models
import hiddenlayer as hl
model = torchvision.models.vgg16()
hl.build_graph(model, torch.zeros([1, 3, 224, 224]))
`
AttributeError: module 'torch.onnx' has no attribute 'OperatorExportTypes'
And I run the code under Ubuntu16.04, pytorch 0.4.0
How to fold when constant is added ?
Hello !
How can I write a proper folding rule for this ?
I don't know how to remove the "Constant" by including it in a folding since it has no parent...
How should I traverse the graph that is built using hidden layer
I would like to traverse the graph generated by hidden layer to investigate the model(any I choose) for architectural changes. How can I do that?
passing model args problem
Hi!
I have a network using convolutionnal and fully connected layers (nothing very new indeed :) ) in torch
img_channels = 5
img_H = 64
img_W = 64
n_batchs = 1
model = Net(img_channels)
imgs = torch.zeros([n_batchs, img_channels,img_H ,img_W])
reds = torch.zeros([n_batchs,1])
When I call it
model(imgs,reds)
I get the correct layer tensor sahpe
input shape: torch.Size([1, 5, 64, 64])
conv0 shape: torch.Size([1, 64, 64, 64])
conv0p shape: torch.Size([1, 64, 32, 32])
i0:START <<<<<<<
Inception x_s1_0 : torch.Size([1, 48, 32, 32])
Inception x_s2_0 : torch.Size([1, 64, 32, 32])
Inception x_s1_2 : torch.Size([1, 48, 32, 32])
Inception x_pool0 : torch.Size([1, 48, 32, 32])
Inception x_s1_1 : torch.Size([1, 48, 32, 32])
Inception x_s2_1 : torch.Size([1, 64, 32, 32])
Inception x_s2_2 : torch.Size([1, 64, 32, 32])
Inception output : torch.Size([1, 240, 32, 32])
i1:START <<<<<<<
Inception x_s1_0 : torch.Size([1, 64, 32, 32])
Inception x_s2_0 : torch.Size([1, 92, 32, 32])
Inception x_s1_2 : torch.Size([1, 64, 32, 32])
Inception x_pool0 : torch.Size([1, 64, 32, 32])
Inception x_s1_1 : torch.Size([1, 64, 32, 32])
Inception x_s2_1 : torch.Size([1, 92, 32, 32])
Inception x_s2_2 : torch.Size([1, 92, 32, 32])
Inception output : torch.Size([1, 340, 32, 32])
i1p shape: torch.Size([1, 340, 16, 16])
i2:START <<<<<<<
Inception x_s1_0 : torch.Size([1, 92, 16, 16])
Inception x_s2_0 : torch.Size([1, 128, 16, 16])
Inception x_s1_2 : torch.Size([1, 92, 16, 16])
Inception x_pool0 : torch.Size([1, 92, 16, 16])
Inception x_s1_1 : torch.Size([1, 92, 16, 16])
Inception x_s2_1 : torch.Size([1, 128, 16, 16])
Inception x_s2_2 : torch.Size([1, 128, 16, 16])
Inception output : torch.Size([1, 476, 16, 16])
i3:START <<<<<<<
Inception x_s1_0 : torch.Size([1, 92, 16, 16])
Inception x_s2_0 : torch.Size([1, 128, 16, 16])
Inception x_s1_2 : torch.Size([1, 92, 16, 16])
Inception x_pool0 : torch.Size([1, 92, 16, 16])
Inception x_s1_1 : torch.Size([1, 92, 16, 16])
Inception x_s2_1 : torch.Size([1, 128, 16, 16])
Inception x_s2_2 : torch.Size([1, 128, 16, 16])
Inception output : torch.Size([1, 476, 16, 16])
i3p shape: torch.Size([1, 476, 8, 8])
i4:START <<<<<<<
Inception x_s1_0 : torch.Size([1, 92, 8, 8])
Inception x_s2_0 : torch.Size([1, 128, 8, 8])
Inception x_s1_2 : torch.Size([1, 92, 8, 8])
Inception x_pool0 : torch.Size([1, 92, 8, 8])
Inception x_s2_2 : torch.Size([1, 128, 8, 8])
Inception output : torch.Size([1, 348, 8, 8])
FC part :START <<<<<<<
flat shape: torch.Size([1, 22272])
concat shape: torch.Size([1, 22273])
fcn_in_features: 22273
fc0 shape: torch.Size([1, 1096])
fc1 shape: torch.Size([1, 1096])
fc2 shape: torch.Size([1, 180])
output shape: torch.Size([1, 180])
But, when I call
hl.build_graph(model,(imgs,reds))
then HL complains
TypeError: forward() takes 2 positional arguments but 3 were given
Any idea?
JE
PS: torch 1.1.0 and a fresh HL install on my mac
BUG report: the stable version on pip has a bug
Hi there~
Your stable version on pip has a bug.
line 324 in graph.py
dot.attr("edge", style="doted",
would cause Error:
Warning: gvrender_set_style: unsupported style doted - ignoring
which was terribly annoying. And caused me lots of time debugging.
We found that you spell the dotted wrongly..., it should be "dotted" instead of "doted". And we also found that you had updated the code into "solid" on github...
We hope you guys update the pip package as soon as possible because the bug is really too funny ^_^.
Unexpected data type for grouping rules? (Codebase merge issue 2 of 3)
shape error
Hi, I met a problem when I want to visualize the darknet:
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print("device = ", device)
model = Darknet("yolo/cfg/yolov3.cfg").to(device)
hl.build_graph(model, torch.zeros([1, 3, 224, 224]).to(device))
RuntimeError: shape '[1, 255, 841]' is invalid for input of size 199920
Thanks.
how can I show graph if don't use jupyter notebook
adaptive_avg_pool2d does not exist
Hi, I met this problem when I want to visualize densenet, error shows below:
c:\python35\lib\site-packages\torch\onnx\utils.py:446: UserWarning: ONNX export failed on ATen operator adaptive_avg_pool2d because torch.onnx.symbolic.adaptive_avg_pool2d does not exist .format(op_name, op_name))
Actually, there is no adaptive_avg_pool2d in my densenet, only nn.AvgPool2d() exists.
Support dict input?
I got a runtime error when I tried to plot a model graph whose input x actually is a dict.
import hiddenlayer as hl
x = getBuff(0)
hl.build_graph(net, x)
So I'm wondering can this plot tool support dict input.
run time error on yolov3
File "getGraph.py", line 13, in
graph = hl.build_graph(model, torch.zeros([1,3, 416, 416])).to(device)
File "/home/tbaggu/Desktop/20-net-works/hiddenlayer/hiddenlayer/graph.py", line 143, in build_graph
import_graph(g, model, args)
File "/home/tbaggu/Desktop/20-net-works/hiddenlayer/hiddenlayer/pytorch_builder.py", line 70, in import_graph
trace, out = torch.jit.get_trace_graph(model, args)
File "/usr/local/lib/python3.5/dist-packages/torch/jit/init.py", line 77, in get_trace_graph
return LegacyTracedModule(f)(*args, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 477, in call
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/jit/init.py", line 109, in forward
out = self.inner(*trace_inputs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 475, in call
result = self._slow_forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 465, in _slow_forward
result = self.forward(*input, **kwargs)
File "/home/tbaggu/Desktop/20-net-works/yolov3/models.py", line 208, in forward
x = module[0](x, img_size)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 475, in call
result = self._slow_forward(*input, **kwargs)
File "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 465, in _slow_forward
result = self.forward(*input, **kwargs)
File "/home/tbaggu/Desktop/20-net-works/yolov3/models.py", line 127, in forward
create_grids(self, img_size, (nx, ny), p.device)
File "/home/tbaggu/Desktop/20-net-works/yolov3/models.py", line 253, in create_grids
self.anchor_vec = self.anchors.to(device) / self.stride
RuntimeError: Expected object of type torch.FloatTensor but found type torch.LongTensor for argument #2 'other
Network is taken from
https://github.com/ultralytics/yolov3/blob/master/train.py
Clusters convolutions with different kernels
I use the newest version bd0d36a. I have a problem that it combines two different convolutions (in this case conv 3x3 > relu and conv 1x1 > relu) in one, incorrect block (conv 3x3 > relu x2).
import torch
from torch import nn
import hiddenlayer as hl
model = nn.Sequential(
nn.Conv2d(8, 8, 3, padding=1),
nn.ReLU(),
nn.Conv2d(8, 8, 1),
nn.ReLU(),
nn.MaxPool2d(2, 2))
hl.build_graph(model, torch.zeros([1, 8, 32, 32]))
When there are different activations, the problem disappears:
model = nn.Sequential(
nn.Conv2d(8, 8, 3, padding=1),
nn.ReLU(),
nn.Conv2d(8, 8, 1),
nn.MaxPool2d(2, 2))
Side notes
- It clusters ReLU with operations (which I like a lot, otherwise there is too much of visual noise), but not other activation functions (sigmoid, tanh, etc); is there some rationale for that? (My preference would be to cluster all reasonable activation functions).
- In general, I am very enthusiastic about neural network visualizations, see my overview: Simple diagrams of convoluted neural networks; I did mine, but for a limited case od purely sequential networks in Keras: https://github.com/stared/keras-sequential-ascii,
x = x.view(x.size(0), 256 * 6 * 6)
？？？？？？
Pointer Network - 2 Inputs issue
When calling:
hl_graph = hl.build_graph(pointer, (inputs, target))
with inputs: tensor([[3, 7, 2, 9, 0, 1, 8, 5, 4, 6]])
and target: tensor([[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]])
I get the following error:
/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py in _slow_forward(self, *input, **kwargs)
463 tracing_state._traced_module_stack.append(self)
464 try:
--> 465 result = self.forward(*input, **kwargs)
466 finally:
467 tracing_state.pop_scope()
<ipython-input-217-844f2c5b2d6a> in forward(self, inputs, target)
80
81
---> 82 return loss / seq_len
RuntimeError: Expected object of type torch.FloatTensor but found type torch.LongTensor for argument #2 'other'
And PointerNetwork is:
class PointerNet(nn.Module):
def __init__(self,
embedding_size,
hidden_size,
seq_len,
n_glimpses,
tanh_exploration,
use_tanh,
use_cuda=USE_CUDA):
super(PointerNet, self).__init__()
self.embedding_size = embedding_size
self.hidden_size = hidden_size
self.n_glimpses = n_glimpses
self.seq_len = seq_len
self.use_cuda = use_cuda
self.embedding = nn.Embedding(seq_len, embedding_size)
self.encoder = nn.LSTM(embedding_size, hidden_size, batch_first=True)
self.decoder = nn.LSTM(embedding_size, hidden_size, batch_first=True)
self.pointer = Attention(hidden_size, use_tanh=use_tanh, C=tanh_exploration, use_cuda=use_cuda)
self.glimpse = Attention(hidden_size, use_tanh=False, use_cuda=use_cuda)
self.decoder_start_input = nn.Parameter(torch.FloatTensor(embedding_size))
self.decoder_start_input.data.uniform_(-(1. / math.sqrt(embedding_size)), 1. / math.sqrt(embedding_size))
self.criterion = nn.CrossEntropyLoss()
def apply_mask_to_logits(self, logits, mask, idxs):
batch_size = logits.size(0)
clone_mask = mask.clone()
if idxs is not None:
clone_mask[[i for i in range(batch_size)], idxs.data] = 1
logits[clone_mask] = -np.inf
return logits, clone_mask
def forward(self, inputs, target):
"""
Args:
inputs: [batch_size x sourceL]
"""
batch_size = inputs.size(0)
seq_len = inputs.size(1)
assert seq_len == self.seq_len
embedded = self.embedding(inputs)
target_embedded = self.embedding(target)
encoder_outputs, (hidden, context) = self.encoder(embedded)
mask = torch.zeros(batch_size, seq_len).byte()
if self.use_cuda:
mask = mask.cuda()
idxs = None
decoder_input = self.decoder_start_input.unsqueeze(0).repeat(batch_size, 1)
loss = 0
for i in range(seq_len):
_, (hidden, context) = self.decoder(decoder_input.unsqueeze(1), (hidden, context))
query = hidden.squeeze(0)
for i in range(self.n_glimpses):
ref, logits = self.glimpse(query, encoder_outputs)
logits, mask = self.apply_mask_to_logits(logits, mask, idxs)
query = torch.bmm(ref, F.softmax(logits).unsqueeze(2)).squeeze(2)
_, logits = self.pointer(query, encoder_outputs)
logits, mask = self.apply_mask_to_logits(logits, mask, idxs)
decoder_input = target_embedded[:,i,:]
loss += self.criterion(logits, target[:,i])
return loss / seq_len
Didn't show strides for Conv in Pytorch 1.1
{'dilations': [1, 1], 'group': 1, 'kernel_shape': [5, 5], 'pads': [2, 2, 2, 2], 'strides': [2, 2]}
the pytorch params show 'strides', (with a s at the end)
code:
if "stride" in self.params: stride = self.params["stride"]
has no 's' at the end, so the graph title will not have the stride info
pip nor conda cannot install hiddenlayer
pip 18.0
conda 4.6.11
ubuntu 16.04
nv-docker2
conda install graphviz python-graphviz
conda list
# Name Version Build Channel
_tflow_select 2.1.0 gpu
absl-py 0.7.1 py36_0 conda-forge
asn1crypto 0.24.0 py36_1003 conda-forge
astor 0.7.1 py_0 conda-forge
backcall 0.1.0 py_0 conda-forge
binutils_impl_linux-64 2.31.1 h6176602_1 conda-forge
binutils_linux-64 2.31.1 h6176602_3 conda-forge
blas 1.0 mkl
bleach 2.1.4 py_1 conda-forge
blinker 1.4 py_1 conda-forge
boto 2.49.0 py36_0
boto3 1.9.53 py_0 conda-forge
botocore 1.12.54 py_0 conda-forge
bz2file 0.98 py_0 conda-forge
bzip2 1.0.6 h470a237_2 conda-forge
c-ares 1.15.0 h14c3975_1001 conda-forge
ca-certificates 2019.3.9 hecc5488_0 conda-forge
cairo 1.16.0 ha4e643d_1000 conda-forge
certifi 2019.3.9 py36_0 conda-forge
cffi 1.11.5 py36h5e8e0c9_1 conda-forge
chardet 3.0.4 py36_1003 conda-forge
cryptography 2.6.1 py36h9d9f1b6_0 conda-forge
cryptography-vectors 2.3.1 py36_1000 conda-forge
cudatoolkit 9.0 h13b8566_0
cudnn 7.3.1 cuda9.0_0
cupti 9.0.176 0
curl 7.64.0 h646f8bb_2 conda-forge
cycler 0.10.0 py_1 conda-forge
dbus 1.13.0 h3a4f0e9_0 conda-forge
decorator 4.3.0 py_0 conda-forge
docutils 0.14 py36_1001 conda-forge
entrypoints 0.2.3 py36_2 conda-forge
expat 2.2.5 hfc679d8_2 conda-forge
fontconfig 2.13.1 h65d0f4c_0 conda-forge
freetype 2.9.1 h6debe1e_4 conda-forge
gast 0.2.2 py_0 conda-forge
gcc_impl_linux-64 7.3.0 habb00fd_1 conda-forge
gcc_linux-64 7.3.0 h553295d_3 conda-forge
gensim 3.5.0 py36_0 conda-forge
gettext 0.19.8.1 h5e8e0c9_1 conda-forge
git 2.21.0 pl526h2882143_0 conda-forge
glib 2.58.3 hf63aee3_1001 conda-forge
gmp 6.1.2 hfc679d8_0 conda-forge
graphite2 1.3.13 hf484d3e_1000 conda-forge
graphviz 2.40.1 h0dab3d1_0 conda-forge
grpcio 1.16.1 py36hf8bcb03_1
gst-plugins-base 1.14.4 hdf3bae2_1001 conda-forge
gstreamer 1.14.4 h66beb1c_1001 conda-forge
gxx_impl_linux-64 7.3.0 hdf63c60_1 conda-forge
gxx_linux-64 7.3.0 h553295d_3 conda-forge
h5py 2.9.0 nompi_py36hf008753_1102 conda-forge
harfbuzz 2.3.1 h6824563_0 conda-forge
hdf5 1.10.4 nompi_h11e915b_1105 conda-forge
html5lib 1.0.1 py_0 conda-forge
icu 58.2 hfc679d8_0 conda-forge
idna 2.7 py36_1002 conda-forge
intel-openmp 2019.1 144
ipykernel 5.0.0 pyh24bf2e0_1 conda-forge
ipython 7.0.1 py36h24bf2e0_0 conda-forge
ipython_genutils 0.2.0 py_1 conda-forge
ipywidgets 7.4.2 py_0 conda-forge
jedi 0.12.1 py36_0 conda-forge
jinja2 2.10 py_1 conda-forge
jmespath 0.9.3 py_1 conda-forge
jpeg 9c h470a237_1 conda-forge
jsonschema 2.6.0 py36_2 conda-forge
jupyter 1.0.0 py_1 conda-forge
jupyter_client 5.2.3 py_1 conda-forge
jupyter_console 5.1.0 py36_0 conda-forge
jupyter_core 4.4.0 py_0 conda-forge
jupyterlab 0.34.12 py36_0 conda-forge
jupyterlab_launcher 0.13.1 py_2 conda-forge
keras 2.1.6 py36_0 conda-forge
keras-applications 1.0.7 py_0 conda-forge
keras-preprocessing 1.0.9 py_0 conda-forge
kiwisolver 1.0.1 py36h2d50403_2 conda-forge
krb5 1.16.3 h05b26f9_1001 conda-forge
libcurl 7.64.0 h541490c_2 conda-forge
libedit 3.1.20170329 haf1bffa_1 conda-forge
libffi 3.2.1 hfc679d8_5 conda-forge
libgcc-ng 7.3.0 hdf63c60_0 conda-forge
libgfortran 3.0.0 1 conda-forge
libgfortran-ng 7.2.0 hdf63c60_3 conda-forge
libgpuarray 0.7.6 h14c3975_1003 conda-forge
libiconv 1.15 h470a237_3 conda-forge
libpng 1.6.35 ha92aebf_2 conda-forge
libprotobuf 3.7.0 h8b12597_2 conda-forge
libsodium 1.0.16 h470a237_1 conda-forge
libssh2 1.8.0 h90d6eec_1004 conda-forge
libstdcxx-ng 7.3.0 hdf63c60_0 conda-forge
libtiff 4.0.9 he6b73bb_2 conda-forge
libtool 2.4.6 h14c3975_1002 conda-forge
libuuid 2.32.1 h470a237_2 conda-forge
libxcb 1.13 h470a237_2 conda-forge
libxml2 2.9.8 h422b904_5 conda-forge
mako 1.0.7 py_1 conda-forge
markdown 2.6.11 py_0 conda-forge
markupsafe 1.0 py36h470a237_1 conda-forge
matplotlib 3.0.3 py36_0 conda-forge
matplotlib-base 3.0.3 py36h167e16e_0 conda-forge
mistune 0.8.3 py36h470a237_2 conda-forge
mkl 2018.0.3 1
mkl_fft 1.0.10 py36_0 conda-forge
mkl_random 1.0.2 py36_0 conda-forge
nb_conda_kernels 2.1.1 py36_1 conda-forge
nbconvert 5.3.1 py_1 conda-forge
nbformat 4.4.0 py_1 conda-forge
ncurses 6.1 hfc679d8_1 conda-forge
ninja 1.8.2 h2d50403_1 conda-forge
nltk 3.2.5 py_0 conda-forge
notebook 5.7.0 py36_0 conda-forge
numpy 1.15.0 py36h1b885b7_0
numpy-base 1.15.0 py36h3dfced4_0
oauthlib 2.1.0 py_0 conda-forge
olefile 0.46 py_0 conda-forge
openblas 0.2.20 8 conda-forge
openssl 1.1.1b h14c3975_1 conda-forge
pandas 0.23.4 py36hf8a1672_0 conda-forge
pandoc 2.3.1 0 conda-forge
pandocfilters 1.4.2 py_1 conda-forge
pango 1.40.14 h4ea9474_1004 conda-forge
parso 0.3.1 py_0 conda-forge
pcre 8.41 hfc679d8_3 conda-forge
perl 5.26.2 h470a237_0 conda-forge
pexpect 4.6.0 py36_0 conda-forge
pickleshare 0.7.5 py36_0 conda-forge
pillow 5.3.0 py36hc736899_0 conda-forge
pip 18.0 py36_1 conda-forge
pixman 0.34.0 h14c3975_1003 conda-forge
prometheus_client 0.3.1 py_1 conda-forge
prompt_toolkit 2.0.5 py_0 conda-forge
protobuf 3.7.0 py36he1b5a44_1 conda-forge
pthread-stubs 0.4 h470a237_1 conda-forge
ptyprocess 0.6.0 py36_1000 conda-forge
pycparser 2.19 py_0 conda-forge
pygments 2.2.0 py_1 conda-forge
pygpu 0.7.6 py36h3010b51_1000 conda-forge
pyjwt 1.6.4 py_0 conda-forge
pyopenssl 18.0.0 py36_1000 conda-forge
pyparsing 2.3.0 py_0 conda-forge
pyqt 4.11.4 py36_3 conda-forge
pysocks 1.6.8 py36_1002 conda-forge
python 3.6.7 h0371630_0
python-crfsuite 0.9.6 py36h2d50403_0 conda-forge
python-dateutil 2.7.3 py_0 conda-forge
python-graphviz 0.10.1 py_0 conda-forge
pytorch 1.0.1 py3.6_cuda9.0.176_cudnn7.4.2_0 pytorch
pytz 2018.5 py_0 conda-forge
pyyaml 5.1 py36h14c3975_0 conda-forge
pyzmq 17.1.2 py36hae99301_0 conda-forge
qt 4.8.7 2
qtconsole 4.4.1 py36_1 conda-forge
readline 7.0 haf1bffa_1 conda-forge
requests 2.20.1 py36_1000 conda-forge
requests-oauthlib 1.0.0 py_1 conda-forge
s3transfer 0.1.13 py36_1001 conda-forge
scikit-learn 0.19.1 py36hedc7406_0
scipy 1.1.0 py36hc49cb51_0
send2trash 1.5.0 py_0 conda-forge
setuptools 40.4.0 py36_1000 conda-forge
simplegeneric 0.8.1 py_1 conda-forge
sip 4.18 py36_1 conda-forge
six 1.11.0 py36_1 conda-forge
smart_open 1.7.1 py_0 conda-forge
sqlite 3.25.2 hb1c47c0_0 conda-forge
tensorboard 1.12.0 py36_1000 conda-forge
tensorboardx 1.6 py_0 conda-forge
tensorflow 1.12.0 gpu_py36he68c306_0
tensorflow-base 1.12.0 gpu_py36h8e0ae2d_0
tensorflow-gpu 1.12.0 h0d30ee6_0
termcolor 1.1.0 py_2 conda-forge
terminado 0.8.1 py36_1 conda-forge
testpath 0.4.1 py36_0 conda-forge
theano 1.0.4 py36hf484d3e_1000 conda-forge
tk 8.6.9 ha92aebf_0 conda-forge
torchvision 0.2.1 py_2 pytorch
tornado 5.1.1 py36h470a237_0 conda-forge
traitlets 4.3.2 py36_0 conda-forge
twython 3.7.0 py_0 conda-forge
urllib3 1.23 py36_1001 conda-forge
wcwidth 0.1.7 py_1 conda-forge
webencodings 0.5.1 py_1 conda-forge
werkzeug 0.14.1 py_0 conda-forge
wheel 0.31.1 py36_1001 conda-forge
widgetsnbextension 3.4.2 py36_0 conda-forge
xorg-kbproto 1.0.7 h14c3975_1002 conda-forge
xorg-libice 1.0.9 h516909a_1004 conda-forge
xorg-libsm 1.2.3 h84519dc_1000 conda-forge
xorg-libx11 1.6.7 h14c3975_1000 conda-forge
xorg-libxau 1.0.8 h470a237_6 conda-forge
xorg-libxdmcp 1.1.2 h470a237_7 conda-forge
xorg-libxext 1.3.4 h516909a_0 conda-forge
xorg-libxpm 3.5.12 h14c3975_1002 conda-forge
xorg-libxrender 0.9.10 h516909a_1002 conda-forge
xorg-libxt 1.1.5 h14c3975_1002 conda-forge
xorg-renderproto 0.11.1 h14c3975_1002 conda-forge
xorg-xextproto 7.3.0 h14c3975_1002 conda-forge
xorg-xproto 7.0.31 h14c3975_1007 conda-forge
xz 5.2.4 h470a237_1 conda-forge
yaml 0.1.7 h14c3975_1001 conda-forge
zeromq 4.2.5 hfc679d8_6 conda-forge
zlib 1.2.11 h470a237_3 conda-forge
(dl) root@dab3b36995ff:/workhere/hiddenlayer#
pip install hiddenlayer
or
clone from github (as shown in the README.md, developer mode)
both results in error:
Exception:
Traceback (most recent call last):
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2869, in _dep_map
return self.__dep_map
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2663, in __getattr__
raise AttributeError(attr)
AttributeError: _DistInfoDistribution__dep_map
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/packaging/requirements.py", line 93, in __init__
req = REQUIREMENT.parseString(requirement_string)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 1632, in parseString
raise exc
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 1622, in parseString
loc, tokens = self._parse( instring, 0 )
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 1379, in _parseNoCache
loc,tokens = self.parseImpl( instring, preloc, doActions )
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 3395, in parseImpl
loc, exprtokens = e._parse( instring, loc, doActions )
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 1383, in _parseNoCache
loc,tokens = self.parseImpl( instring, preloc, doActions )
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pyparsing.py", line 3183, in parseImpl
raise ParseException(instring, loc, self.errmsg, self)
pip._vendor.pyparsing.ParseException: Expected stringEnd (at char 33), (line:1, col:34)
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2949, in __init__
super(Requirement, self).__init__(requirement_string)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/packaging/requirements.py", line 97, in __init__
requirement_string[e.loc:e.loc + 8]))
pip._vendor.packaging.requirements.InvalidRequirement: Invalid requirement, parse error at "'; extra '"
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_internal/basecommand.py", line 141, in main
status = self.run(options, args)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_internal/commands/install.py", line 330, in run
self._warn_about_conflicts(to_install)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_internal/commands/install.py", line 456, in _warn_about_conflicts
package_set, _dep_info = check_install_conflicts(to_install)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_internal/operations/check.py", line 98, in check_install_conflicts
package_set = create_package_set_from_installed()
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_internal/operations/check.py", line 41, in create_package_set_from_installed
package_set[name] = PackageDetails(dist.version, dist.requires())
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2607, in requires
dm = self._dep_map
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2871, in _dep_map
self.__dep_map = self._compute_dependencies()
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2881, in _compute_dependencies
reqs.extend(parse_requirements(req))
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2942, in parse_requirements
yield Requirement(line)
File "/root/anaconda3/envs/dl/lib/python3.6/site-packages/pip/_vendor/pkg_resources/__init__.py", line 2951, in __init__
raise RequirementParseError(str(e))
pip._vendor.pkg_resources.RequirementParseError: Invalid requirement, parse error at "'; extra '"
module 'torch.onnx' has no attribute 'OperatorExportTypes'
I run this code in Jupyter Notebook，but one error occurs:
`import torch
import torchvision.models
import hiddenlayer as hl
VGG16 with BatchNorm
model = torchvision.models.vgg16()
Build HiddenLayer graph
Jupyter Notebook renders it automatically
hl.build_graph(model, torch.zeros([1, 3, 224, 224]))`
AttributeError: module 'torch.onnx' has no attribute 'OperatorExportTypes'
And I run the code under Ubuntu16.04, pytorch 0.4.0
TypeError: expected string or bytes-like object
Get error then trying to plot this model
https://drive.google.com/open?id=1QpcZb2nnzNmXAEm-DTBkRW2qKAb7hRq5
Code:
https://colab.research.google.com/drive/1UR07U-9g115eLr2q83oojKThXqGzBtPY
Keras output section of graph is very complex
The output section of every Keras graph is complex, and has a long chain of TF nodes that uses up a lot of screen space.
Request some way to Prune everything downstream of the "beginning of the outputs", and rename the end block "OUTPUT". Graphs with multiple outputs might have "OUTPUT #1" etc.
Possible style misspelling
hiddenlayer/hiddenlayer/graph.py
Line 324 in 3588dbf
Should this be dotted maybe? When I build_graph with a PyTorch model, I get
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
Warning: gvrender_set_style: unsupported style doted - ignoring
RNN can not display correctly
- Here is a simple example of LSTM neural network.
We need hl.transforms.Rename() to rename the RNN node.
tsfm = [hl.transforms.Rename(op='prim::PythonOp', to = 'LSTM')]
TypeError: 'Metric' object does not support indexing
Hi, I installed hiddenlayer and then tried to run the demo "history_canvas.py". But there is something wrong. The error message is "Step 0: loss: 1.042669959384648 accuracy: 0.06743948517530154
Traceback (most recent call last):
File "./demos/history_canvas.py", line 43, in
c.draw_plot(h["loss"], h["accuracy"])
File "/home/yangbiao/hiddenlayer-master/hiddenlayer/canvas.py", line 153, in wrapper
self.render()
File "/home/yangbiao/hiddenlayer-master/hiddenlayer/canvas.py", line 133, in render
getattr(self, method)(*c[1], **c[2])
File "/home/yangbiao/hiddenlayer-master/hiddenlayer/canvas.py", line 180, in draw_plot
label = labels[i] if labels else m.name
TypeError: 'Metric' object does not support indexing."
I used python 3.5.2 and did not use jupter note. I'm not sure whether the mistake is caused by the python version? Please give me some advice about it. Thx.
"torch._C.Value has no attribute 'uniqueName'" Error running with PyTorch 1.2
PyTorch Version: '1.2.0a'
Python: 3.6.8
Exception has occurred: AttributeError 'torch._C.Value' object has no attribute 'uniqueName' File "hiddenlayer/hiddenlayer/pytorch_builder.py", line 45, in <listcomp> return node.scopeName() + "/outputs/" + "/".join([o.uniqueName() for o in node.outputs()]) File "hiddenlayer/hiddenlayer/pytorch_builder.py", line 45, in pytorch_id return node.scopeName() + "/outputs/" + "/".join([o.uniqueName() for o in node.outputs()]) File "hiddenlayer/hiddenlayer/pytorch_builder.py", line 90, in import_graph hl_node = Node(uid=pytorch_id(torch_node), name=None, op=op, File "hiddenlayer/hiddenlayer/graph.py", line 143, in build_graph import_graph(g, model, args) File "visualizer.py", line 20, in <module> graph = hl.build_graph(model, input)
Works well with older version of PyTorch (0.4.1).
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.