未验证 提交 5f1b3ac7 编写于 作者: Z Zihao Mu 提交者: GitHub

Merge pull request #909 from zihaomu:tf_biasadd_bug

DNN: Test data of biasadd in tf_importer

* Add files via upload

* Update generate_tf_models.py
上级 9c9ae8f7
......@@ -1027,6 +1027,11 @@ concat = tf.concat([resized1, biased], 3)
# writeBlob(blob, 'resize_concat_optimization_in')
save(inp, concat, 'resize_concat_optimization', optimize=False, is_gen_data=False)
################################################################################
inp = tf.placeholder(tf.float32, [1, 2, 3, 4], 'input')
sub = inp - 3.0
sub = 4.0 + sub
save(inp, sub, prefix + 'bias_add_1', optimize=False)
################################################################################
# Uncomment to print the final graph.
# with tf.gfile.FastGFile('fused_batch_norm_net.pb', 'rb') as f:
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册