F tensorflow/python/lib/core/bfloat16.cc:664] Check failed: PyBfloat16_Type.tp_base != nullptr Aborted

[monajalal]

So when running the test.py code, I get this error:

    [jalal@goku pose-tensorflow]$ TF_CUDNN_USE_AUTOTUNE=0 python demo/demo_multiperson.py
    RuntimeError: module compiled against API version 0xc but this version of numpy is 0xb
    ImportError: numpy.core.multiarray failed to import
    ImportError: numpy.core.umath failed to import
    ImportError: numpy.core.umath failed to import
    2018-04-08 17:09:02.321666: F tensorflow/python/lib/core/bfloat16.cc:664] Check failed: PyBfloat16_Type.tp_base != nullptr 
    Aborted

The test.py code is:

  import argparse
   import logging
   import os

   import numpy as np
   import scipy.io
   import scipy.ndimage
   
   from config import load_config
   from dataset.factory import create as create_dataset
   from dataset.pose_dataset import Batch
   from nnet.predict import setup_pose_prediction, extract_cnn_output, argmax_pose_predict
   from util import visualize
   
   
   def test_net(visualise, cache_scoremaps):
       logging.basicConfig(level=logging.INFO)
   
       cfg = load_config()
       dataset = create_dataset(cfg)
       dataset.set_shuffle(False)
       dataset.set_test_mode(True)
   
       sess, inputs, outputs = setup_pose_prediction(cfg)
   
       if cache_scoremaps:
           out_dir = cfg.scoremap_dir
           if not os.path.exists(out_dir):
               os.makedirs(out_dir)
   
       num_images = dataset.num_images
       predictions = np.zeros((num_images,), dtype=np.object)
   
       for k in range(num_images):
           print('processing image {}/{}'.format(k, num_images-1))
   
           batch = dataset.next_batch()
   
           outputs_np = sess.run(outputs, feed_dict={inputs: batch[Batch.inputs]})
   
           scmap, locref, pairwise_diff = extract_cnn_output(outputs_np, cfg)
   
           pose = argmax_pose_predict(scmap, locref, cfg.stride)
   
           pose_refscale = np.copy(pose)
           pose_refscale[:, 0:2] /= cfg.global_scale
           predictions[k] = pose_refscale
   
           if visualise:
               img = np.squeeze(batch[Batch.inputs]).astype('uint8')
               visualize.show_heatmaps(cfg, img, scmap, pose)
               visualize.waitforbuttonpress()
   
           if cache_scoremaps:
               base = os.path.basename(batch[Batch.data_item].im_path)
               raw_name = os.path.splitext(base)[0]
               out_fn = os.path.join(out_dir, raw_name + '.mat')
               scipy.io.savemat(out_fn, mdict={'scoremaps': scmap.astype('float32')})
   
               out_fn = os.path.join(out_dir, raw_name + '_locreg' + '.mat')
               if cfg.location_refinement:
                   scipy.io.savemat(out_fn, mdict={'locreg_pred': locref.astype('float32')})
   
       scipy.io.savemat('predictions.mat', mdict={'joints': predictions})
   
       sess.close()
   
   
   if __name__ == '__main__':
       parser = argparse.ArgumentParser()
       parser.add_argument('--novis', default=False, action='store_true')
       parser.add_argument('--cache', default=False, action='store_true')
       args, unparsed = parser.parse_known_args()
   
       test_net(not args.novis, args.cache)

I have the followings:

  $ conda list tensorflow
   # packages in environment at /scratch/sjn/anaconda:
   #
   tensorflow                1.5.0                    py36_0    conda-forge
   tensorflow-gpu            1.3.0                         0  
   tensorflow-gpu-base       1.3.0           py36cuda8.0cudnn6.0_1  
   tensorflow-tensorboard    0.1.5                    py36_0  

and

$ conda list numpy
  # packages in environment at /scratch/sjn/anaconda:
  #
  msgpack-numpy             0.4.1                     <pip>
  numpy                     1.13.3          py36_blas_openblas_201  [blas_openblas]  conda-forge
  numpydoc                  0.7.0            py36h18f165f_0  

[monajalal]

Installed tensorflow 1.4.0 and the issue was resolved!

[Huixxi]

conda install tensorflow
I solved this issue with it.

[elan69]

still dosnt work @Huixxi

[lidejian]

@elanChezhianMuthuvel it needs numpy>=1.14,so pip install numpy==1.14 and it work!

[foxjr]

I'm using a custom TF 1.10.0 configuration that doesn't use AVX instructions. I got this error when using the numpy version that accompanied my custom configuration (1.14.1); however, upgrading to latest (1.16.x) resolved the issue, despite a prompt that it could be incompatible.

[weltonmattos]

I'm using a custom TF 1.10.0 configuration that doesn't use AVX instructions. I got this error when using the numpy version that accompanied my custom configuration (1.14.1); however, upgrading to latest (1.16.x) resolved the issue, despite a prompt that it could be incompatible.

i solved by uninstalling current version from numpy and by reinstalling with the last version.
pip uninstall numpy and pip install numpy

[dinakarmaurya-innova]

above problem was same with me - nice answer by weltonmattos

pip uninstall numpy and pip install numpy
above will fix issue.
Thank you.

[wzelang]

tensorflow 1.13 needs numpy >=1.16, update numpy from 1.14 to 1.16 works for me.