release v0.2.0

make --reference_path not a required input in extract and call_mods module

handle gz file in call_freq script

release v0.1.10

make sure results of each read be written together in call_mods' output

release v0.1.9

enable multi-class (>=3) training/predicting,

fix bug of extrating contig name from fast5s

release v0.1.8

modify denoise module
fix success_file bug
update README

release v0.1.7

Prevent Queue.qsize() from raising NotImplementedError on Mac OS X (github: vterron/lemon@9ca6b4b)
covert raw signals to pA values before normalization in extract_feature module
add denoise module
add module-chosen options of model (rnn/base/cnn), re-train human model

release v0.1.6

add option –positions in extract_features module, add option/function of binary_format feature file to speed up training

release v0.1.5

normalize probs before output

release v0.1.4

change the loss function to weighted_cross_entropy_with_logits, allow training using unbalanced samples.

release v0.1.3

fix the deadlock issue in multiprocessing

release v0.1.2

this is for adding the missing MANIFEST.in file, no bugs fix or imporvements.