8000 Placeholder default size · Issue #17 · SymJAX/SymJAX · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Placeholder default size #17

Open
Koldh opened this issue Sep 4, 2020 · 0 comments
Open

Placeholder default size #17

Koldh opened this issue Sep 4, 2020 · 0 comments

Comments

@Koldh
Copy link
Contributor
Koldh commented Sep 4, 2020

Problem with: "indices = T.Placeholder((0,),'int32')", it works if it is replace with: "indices= T.Placeholder((batch_size,),'int32')

import symjax as sj
import symjax.tensor as T
import numpy as np
import scipy
import matplotlib.pyplot as plt
from sklearn.cluster.k_means_ import *
import scipy.sparse as sp
from sklearn.cluster.k_means_ import k_init
from sklearn.cluster.k_means import _kmeans_single_lloyd
from sklearn.cluster import MiniBatchKMeans

class RAI_KMeans:
def init(self,x_shape,n_data_train,n_data_test,batch_size,n_clusters,n_landmarks):

    #placeholder
    data      = T.Placeholder((batch_size,*x_shape),'float32')
    indices   = T.Placeholder((0,),'int32')
    #trainable
    centroids = T.Variable(sj.nn.initializers.glorot_uniform((n_clusters,*x_shape)))
    theta_aff_train = T.Variable(np.tile(np.array([1,0,0,0,1,0]),(n_data_train,1)))
    theta_diff_train = T.Variable(np.zeros((n_data_train,2*n_landmarks)))
    theta_aff_test = T.Variable(T.Variable(np.tile(np.array([1,0,0,0,1,0]),(n_data_train,1))))
    theta_diff_test = T.Variable(np.z
5755
eros((n_data_test,n_clusters,2*n_landmarks)))



    theta_aff_train_ind = theta_aff_train[indices]
    theta_diff_train_ind = theta_diff_train[indices]
    theta_aff_test_ind = theta_aff_test[indices]
    theta_diff_test_ind = theta_aff_test[indices]

    data_train_aff    = T.interpolation.affine_transform(data,theta_aff_train_ind)
    print(data_train_aff.shape.get())
    data_train_diff   = T.interpolation.thin_plate_spline(data_train_aff,theta_diff_train_ind)
    self.get_data_aff = sj.function(data,indices,outputs=data_train_diff)

data = sj.data.mnist()["train_set/images"].astype('float32')
x_shape = data.shape[1:]
n_data_train = data.shape[0]
batch_size = 10
ind_train = np.random.permutation(data.shape[0])[:batch_size].astype('int32')

print(x_shape)

data_train = data[ind_train]

test = RAI_KMeans(x_shape,n_data_train,10,batch_size,5,4)
image = test.get_data_aff(data_train,ind_train)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@Koldh
0