akar43 / lsm Goto Github PK
View Code? Open in Web Editor NEWCode for Learnt Stereo Machines based on the NIPS 2017 paper
License: MIT License
Code for Learnt Stereo Machines based on the NIPS 2017 paper
License: MIT License
I am very interested in the work, and want to test the code. However, it seems that the links to the pretrained models are invalid. Anyone can help?
Hi,
Seems the data is not in the page anymore.
--2019-04-30 00:33:11-- http://people.eecs.berkeley.edu/~akar/lsm/shapenet_release.tar.gz
Resolving people.eecs.berkeley.edu (people.eecs.berkeley.edu)... 128.32.189.73
Connecting to people.eecs.berkeley.edu (people.eecs.berkeley.edu)|128.32.189.73|:80... connected.
HTTP request sent, awaiting response... 301 Moved Permanently
Location: https://abhishekkar.info/lsm/shapenet_release.tar.gz [following]
--2019-04-30 00:33:11-- https://abhishekkar.info/lsm/shapenet_release.tar.gz
Resolving abhishekkar.info (abhishekkar.info)... 185.199.108.153, 185.199.110.153
Connecting to abhishekkar.info (abhishekkar.info)|185.199.108.153|:443... connected.
HTTP request sent, awaiting response... 404 Not Found
2019-04-30 00:33:11 ERROR 404: Not Found.
Hi there.
The dataset url http://people.eecs.berkeley.edu/~akar/lsm/shapenet_release.tar.gz is no longer available.
Is there any other alternative ways I can get the camera and depth data?
Hello! Thank you for releasing the code. I wonder if it is possible to evaluate the D-LSM in single-image mode? In the paper, it states that LSM supports single-image mode, but it seems that it is only related to V-LSM.
hi, when I run lsm, it shows
ImportError Traceback (most recent call last)
in ()
----> 1 from shapenet import ShapeNet
2 # Read data
3 dset = ShapeNet(im_dir=im_dir, split_file=os.path.join(SAMPLE_DIR, 'splits_sample.json'), rng_seed=1)
4 test_mids = dset.get_smids('test')
/home/cloud/Downloads/lsm-master/shapenet.py in ()
8
9 from config import SHAPENET_IM
---> 10 from loader import read_camera, read_depth, read_im, read_quat, read_vol
11
12
/home/cloud/Downloads/lsm-master/loader.py in ()
3 from scipy.io import loadmat
4 from skimage.io import imread
----> 5 from skimage.transform import resize
6
7
/home/cloud/Downloads/lsm-master/env/lib/python2.7/site-packages/skimage/transform/init.py in ()
----> 1 from .hough_transform import (hough_line, hough_line_peaks,
2 probabilistic_hough_line, hough_circle,
3 hough_circle_peaks, hough_ellipse)
4 from .radon_transform import (radon, iradon, iradon_sart,
5 order_angles_golden_ratio)
/home/cloud/Downloads/lsm-master/env/lib/python2.7/site-packages/skimage/transform/hough_transform.py in ()
2 from scipy import ndimage
3 from .. import measure
----> 4 from ._hough_transform import (_hough_circle,
5 _hough_ellipse,
6 _hough_line,
ImportError: cannot import name _hough_circle
Hi,
The url seems to be invalidated for file get_models.sh,it possible to reupload the file, please. @akar43
http://people.eecs.berkeley.edu/~akar/lsm/models_lsm_v1_small.tar.gz
Thanks.
How can one save a D-LSM visualization as a 3d model file, like *.ply?
Hello! I run the demo_vlsm.ipynb in recent days.
Then I found some bug make the code can not work.
That bug is in the "proj_splat()", some times, the projection result may out of the 2D feature map.
For example, the feature map has a size 112*112, the index in each dim is 0~111, but some 3D point project to the 2D feature map, the result will be (112,112), then the model can not work.
current time, I modify the _get_gather_inds(x, y) to solve the problem, the code shows as below :
def _get_gather_inds(x, y):
temp = tf.reshape(tf.stack([im_ind, y, x], axis=2), [-1, 3])
correct_tensor = tf.zeros(temp.get_shape(),dtype=tf.dtypes.int32) + 111
result = tf.where(tf.less(temp, 112), temp , correct_tensor)
return result
Hi, @akar43 ,
When I run the following cell in demo_vlsm.ipynb
,
# Setup TF graph and initialize VLSM model
tf.reset_default_graph()
# Change the ims_per_model to run on different number of views
bs, ims_per_model = 1, 4
ckpt = 'mvnet-100000'
net = MVNet(vmin=-0.5, vmax=0.5, vox_bs=bs,
im_bs=ims_per_model, grid_size=args.nvox,
im_h=args.im_h, im_w=args.im_w,
norm=args.norm, mode="TEST")
net = model_vlsm(net, im_nets[args.im_net], grid_nets[args.grid_net],
conv_rnns[args.rnn])
vars_restore = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='MVNet')
sess = tf.InteractiveSession(config=get_session_config())
saver = tf.train.Saver(var_list=vars_restore)
saver.restore(sess, os.path.join(log_dir, ckpt))
I got the this running error (details log):
ValueError: Dimension must be 6 but is 3 for 'MVNet/transpose' (op: 'Transpose') with input shapes: [1,4,32,32,32,33], [3].
Any suggestion to fix this issue?
My system is Tensorflow 1.0.1 (python 2.7). Is the tensorflow version compatibility issue? What version of tensorflow is required to run your package?
Thanks!
Hi,
I downloaded the data that has been provided. Thanks for that!
But, I wanted images at more azimuths and elevations and with closer camera parameters. So, could you provide the rendering code that you used to generate the renders?
Thanks
Hello! Thank you for sharing your research.
When I use the python 3.5 to run your code, It has some error come from python version.
EX: AttributeError: 'dict' object has no attribute 'iteritems' (in py3.5 should be 'items')
I think is better that you post your python version in readme.md or some file, It can help people more easy to update the code to their working environment.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.