search

haotianteng / Chiron

A basecaller for Oxford Nanopore Technologies' sequencers
Other
122 stars 53 forks source link

Error when training model using raw signals #89

Closed ardakdemir closed 5 years ago

ardakdemir commented 5 years ago

I am trying to reproduce the results obtained by training a model on the same dataset with .signal and .label files. When I run the code I get the following error related to tensorflow:

Traceback (most recent call last): File "chiron/chiron_rcnn_train.py", line 229, in run(args) File "chiron/chiron_rcnn_train.py", line 176, in run train() File "chiron/chiron_rcnn_train.py", line 80, in train net = compile_train_graph(config,FLAGS) File "chiron/chiron_rcnn_train.py", line 58, in compile_traingraph net.error,net.errors,net.y = model.prediction(net.logits, net.seq_length, net.y) File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 439, in iter "Tensor objects are not iterable when eager execution is not "

python version : 2.7.12 tensorflow version : 1.8.0

haotianteng commented 5 years ago

The current version of Chiron use the TFRECORD format, instead of the original .signal and .label format. If you want to reproduce the result, use chiron v0.3 and tf-0.1.

ardakdemir commented 5 years ago

Thanks a lot for your quick reply.

Is there a way to generate the tfrecords from the .signal and .label files rather than the fasta files?