Correct path and name of pretrained model file

Resolves #1
Also updates example model to rgrgr network.

scripts/example_training.sh

@@ -5,7 +5,7 @@ then
 fi
 
 
-git clone https://github.com/nanoporetech/sloika
+git clone --depth 1 https://github.com/nanoporetech/sloika
 (
 	cd sloika
 	make cleanDevEnv
@@ -19,22 +19,22 @@ THEANO_FLAGS_GPU=device=gpu,floatX=float32,mode=FAST_RUN,blas.ldflags='-lblas',s
 
 
 #  Generate reference sequences -- replace with your own method
-sloika/bin/extract_reference.py reads references.fa
+extract_reference.py reads references.fa
 
 #  Map reads using model -- takes a few hours
-THEANO_FLAGS=${THEANO_FLAGS_CPU} sloika/bin/chunkify.py raw_remap --jobs ${NCPU} --chunk_len 4000 --downsample_factor 5 --output_strand_list unfiltered_strands.txt  reads remapped_unfiltered.hdf5 sloika/models/retrained.pkl references.fa
+THEANO_FLAGS=${THEANO_FLAGS_CPU} chunkify.py raw_remap --jobs ${NCPU} --chunk_len 4000 --downsample_factor 5 --output_strand_list unfiltered_strands.txt  reads remapped_unfiltered.hdf5 sloika/models/pretrained.pkl references.fa
 
 #  Filter reads -- criterion from distribution of mapping scores,  coverage and proportion of stays
 ( head -n 1 unfiltered_strands.txt ; cat unfiltered_strands.txt | awk '$3 > 0.5 && $3 < 1.2 && ($7 - $6) > 0.95 * $5 && $5 / ($7 - $6 + $5) < 0.55' ) > filtered_strand_list.txt
 
 #  Remap selected reads -- takes a few hours
-THEANO_FLAGS=${THEANO_FLAGS_CPU} sloika/bin/chunkify.py raw_remap --jobs ${NCPU} --chunk_len 4000 --downsample_factor 5 --input_strand_list filtered_strand_list.txt --output_strand_list filtered_strands.txt reads remapped_filtered.hdf5 model/baseline.pkl references.fa
+THEANO_FLAGS=${THEANO_FLAGS_CPU} chunkify.py raw_remap --jobs ${NCPU} --chunk_len 4000 --downsample_factor 5 --input_strand_list filtered_strand_list.txt --output_strand_list filtered_strands.txt reads remapped_filtered.hdf5 sloika/models/pretrained.pkl references.fa
 
 #  Train a model
-THEANO_FLAGS=${THEANO_FLAGS_GPU} sloika/bin/train_network.py raw  --min_prob 1e-5 sloika/models/raw_1.00_rGr.py training remapped_filtered.hdf5
+THEANO_FLAGS=${THEANO_FLAGS_GPU} train_network.py raw  --min_prob 1e-5 sloika/models/raw_0.98_rgrgr.py training remapped_filtered.hdf5
 
 #  Convert model to CPU
 THEANO_FLAGS=${THEANO_FLAGS_GPU} sloika/misc/model_convert.py --target cpu training/model_final.pkl training/model_final_cpu.pkl
 
 #  Basecall (slowly)
-THEANO_FLAGS=${THEANO_FLAGS_CPU} sloika/bin/basecall_network.py raw --jobs ${NCPU} training/model_final_cpu.pkl test_reads > basecalls.fa
+THEANO_FLAGS=${THEANO_FLAGS_CPU} basecall_network.py raw --jobs ${NCPU} training/model_final_cpu.pkl test_reads > basecalls.fa