I can't run tf.random.normal([1000, 1000])

TensorFlow General Discussion
,
1

When I run the following code :

import os
os.environ['OPENBLAS_NUM_THREADS'] = '1'
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'
from flask import Flask, jsonify
import numpy as np
import tensorflow as tf

app = Flask(__name__)

@app.route('/', methods=['GET'])
def sum_random():
    result = tf.reduce_sum(tf.random.normal([1000, 1000]))
    return jsonify({'result': result.numpy()})

if __name__ == '__main__':
    app.run(debug=True)

I got the following error message:
2024-04-25 10:52:25.965625: F external/local_tsl/tsl/platform/default/env.cc:74] Check failed: ret == 0 (11 vs. 0)Thread tf_numa_-1_Eigen creation via pthread_create() failed.
[UID:1484][2325686] Child process with pid: 2327985 was killed by signal: 6, core dumped: yes
2024-04-25 10:52:26.396146: F external/local_tsl/tsl/platform/default/env.cc:74] Check failed: ret == 0 (11 vs. 0)Thread tf_numa_-1_Eigen creation via pthread_create() failed.
[UID:1484][2325686] Child process with pid: 2328392 was killed by signal: 6, core dumped: yes
2024-04-25 10:52:26.786404: F external/local_tsl/tsl/platform/default/env.cc:74] Check failed: ret == 0 (11 vs. 0)Thread tf_numa_-1_Eigen creation via pthread_create() failed.
[UID:1484][2325686] Child process with pid: 2328774 was killed by signal: 6, core dumped: yes

The following is my specification:
Python 3.11.8
Tensorflow 2.16.1
Flask 3.0.3
Numpy 1.26.4

How do i resolve this issue ?

3

Hi @itoy827, I suspect that this error might be caused due to the memory constraint. Could you please try with the lesser shape while defining the tf.random.normal. Thank You