diff --git a/local/classification_kmeans.py b/local/classification_kmeans.py
index b8265bda098237e27fef9b3313ef53c3f02419b7..70227fb902e00ec6db664119a6f4c4ba32884d0d 100644
--- a/local/classification_kmeans.py
+++ b/local/classification_kmeans.py
@@ -115,7 +115,7 @@ def main(X_path, y_path, pcaA_path, pcab_path, out_path, json_path, NC):
kmeans = faiss.Kmeans(
d,
int(NC),
- niter=10,
+ niter=100,
verbose=True,
gpu=False,
)
diff --git a/local/codebook_freqs.py b/local/codebook_freqs.py
index 0b7e2eaadaf50809f435866d7bf283ed43992462..9b022d1797afc4609d7979ec5a5d64a795c74fed 100644
--- a/local/codebook_freqs.py
+++ b/local/codebook_freqs.py
@@ -35,12 +35,14 @@ def calc_codebook_indexes(audio_path, freq, N):
N = N + Nwav
print('feature vectors: {}/{} (file/all)'.format(Nwav, N))
else:
+ Nwav = 0
print('WARNING: Input size of file is {} (smaller than Kernel size), skip ...'.format(len(x)))
return freq, Nwav
def main(exp_path, lst_path, json_path, model_path, VERBOSE):
# set model and device global
global model, device
+ #device = torch.device('cpu')
device = torch.device('cuda')
# load existing model
model, cfg, task = fairseq.checkpoint_utils.load_model_ensemble_and_task([model_path])
@@ -72,13 +74,14 @@ def main(exp_path, lst_path, json_path, model_path, VERBOSE):
rows = tsv.readlines()
for row in rows:
uttID, audio_path = row.split()
- corpus_match = ('_').join(audio_path.split('/')[1].split('_')[1:]) # DATA/data_corpus_speakingstyle/spk/*wav
- spk_match = audio_path.split('/')[2] # DATA/data_corpus_speakingstyle/spk/*wav
+ corpus_match = ('_').join(audio_path.split('/')[2].split('_')[1:]) # DATA/expname/data_corpus_speakingstyle/spk/*wav
+ spk_match = audio_path.split('/')[3] # DATA/expname/data_corpus_speakingstyle/spk/*wav
if corpus == corpus_match and spk == spk_match:
print('\nread wav-file {}'.format(audio_path))
freq, Nwav = calc_codebook_indexes(audio_path, freq, N)
- N = N + Nwav
- processed_files.append(audio_path)
+ if Nwav != 0:
+ N = N + Nwav
+ processed_files.append(audio_path)
print(f'(DONE) Found {N} observations for speaker {spk} ...')
# combine frequencies
freqN_vec = np.zeros((1, 320**2), dtype=np.float32)
diff --git a/local/plot_pca_similarities.py b/local/plot_pca_similarities.py
index 55e6153301fcf8d3c9bd456096901d219dc8c9d6..07f56fe56081b75b27bed98a90bcf95f1fe6cff5 100644
--- a/local/plot_pca_similarities.py
+++ b/local/plot_pca_similarities.py
@@ -30,6 +30,7 @@ def main(X_path, y_path, pcaA_path, pcab_path, out_path, json_path):
with open(json_path, 'r') as f:
corpora = json.load(f)
+ smID = X_path.split('/')[-1].replace('.npy','')
X = np.load(X_path)
spks = np.load(y_path, allow_pickle=True)
@@ -52,7 +53,7 @@ def main(X_path, y_path, pcaA_path, pcab_path, out_path, json_path):
plt.yticks(range(len(labels)), labels);
cbar = fig.colorbar(cax, ticks=[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, .8, .9, 1])
cbar.ax.tick_params(labelsize=25)
- fig_path = os.path.join(out_path,'{}.png'.format(X_path.split('/')[-1].replace('.npy','')))
+ fig_path = os.path.join(out_path,f'{smID}.png')
print('\nsave similarity matrix to {}'.format(fig_path))
plt.savefig(fig_path)
@@ -88,7 +89,7 @@ def main(X_path, y_path, pcaA_path, pcab_path, out_path, json_path):
# Put a legend to the right of the current axis
ax.legend(loc='center left', bbox_to_anchor=(1.175, .5))
# Save the plot
- fig_path = os.path.join(out_path,'scatter3D_{}.png'.format(X_path.split('/')[-1].replace('.npy','')))
+ fig_path = os.path.join(out_path,f'scatter_{smID}_PCA1_PCA2_PCA3.png')
print('\nsave 3D scatter plot {}'.format(fig_path))
plt.savefig(fig_path)
@@ -113,7 +114,8 @@ def main(X_path, y_path, pcaA_path, pcab_path, out_path, json_path):
plt.ylabel(f'PCA{str(proj[1]+1)}')
plt.grid()
plt.tight_layout()
- fig_path = os.path.join(out_path,'scatter_proj{}_{}.png'.format(''.join([str(proj[0]),str(proj[1])]), X_path.split('/')[-1].replace('.npy','')))
+ pcaID1, pcaID2 = str(proj[0]+1), str(proj[1]+1)
+ fig_path = os.path.join(out_path,f'scatter_{smID}_PCA{pcaID1}_PCA{pcaID2}.png')
print('save 2D scatter plot {}'.format(fig_path))
plt.savefig(fig_path)
diff --git a/run.sh b/run.sh
index 5877ed72c804b72e8fbca80e71c6fc1520cdecd3..74f1aee5d206b79b2e55d24489c0a7c2af284958 100755
--- a/run.sh
+++ b/run.sh
@@ -8,14 +8,15 @@ set -e -o pipefail
. path.sh
. conda.sh
-if [[ $# -eq 1 ]] ; then
- echo 'ERROR: this run-script requires two arguments: DATA=? stage=?'
+if [[ $# -eq 2 ]] ; then
+ echo 'ERROR: this run-script requires two arguments: expdata=? expname=? stage=?'
exit 1
fi
-## runDATA/STAGE
-runDATA=$1
-stage=$2
+## expname/STAGE
+expdata=$1
+expname=$2
+stage=$3
VERBOSE=1 # write logs for codebook frequency extraction?
printf "\n### STAGE ###\n"
printf "stage: %d\n" $stage
@@ -23,7 +24,7 @@ printf "### STAGE ###\n"
## DIRS/PATHS
model_path=model/xlsr_53_56k_new.pt
-exp_dir=exp_$runDATA
+exp_dir=exp_$expname
## STAGE 0: DELETE AND RUN ALL STAGES
if [ $stage == 0 ]; then
@@ -34,7 +35,8 @@ fi
## print:
printf "\nCWD: %s" "$CWD"
printf "\nFAIRSEQ: %s" "$FAIRSEQ"
-printf "\nrunDATA: %s" "$runDATA"
+printf "\nexpname: %s" "$expname"
+printf "\nexpdata: %s" "$expdata"
printf "\nmodel_path: %s" "$model_path"
printf "\nexp_dir: %s\n\n" "$exp_dir"
@@ -50,22 +52,22 @@ mkdir -p $exp_dir/numpy/pca
## PREPARE DATA
if [ $stage == 1 ] || [ $stage == 0 ]; then
printf "\n... Prepare data (*lst and *json) ...\n"
- python3 local/prepare_data.py --output_lst_path $exp_dir/data/${runDATA}.lst \
- --output_json_path $exp_dir/data/${runDATA}.json \
- --DATA_dir ${runDATA}
+ python3 local/prepare_data.py --output_lst_path $exp_dir/data/${expname}.lst \
+ --output_json_path $exp_dir/data/${expname}.json \
+ --DATA_dir ${expdata}
fi
## COUNT CODEBOOK USAGE
if [ $stage == 2 ] || [ $stage == 0 ]; then
printf "\n... Count frequencies of codebooks ...\n"
python3 local/codebook_freqs.py $exp_dir \
- $exp_dir/data/${runDATA}.lst \
- $exp_dir/data/${runDATA}.json \
+ $exp_dir/data/${expname}.lst \
+ $exp_dir/data/${expname}.json \
$model_path \
$VERBOSE
printf "\n... Combine Arrays ... \n"
python3 local/codebook_combine_arrays.py $exp_dir/numpy \
- $exp_dir/data/${runDATA}.json
+ $exp_dir/data/${expname}.json
fi
## CALCULATE SIMILARITY MATRIX
@@ -75,7 +77,7 @@ if [ $stage == 3 ] || [ $stage == 0 ]; then
$exp_dir/numpy/splits_freqs.npy \
$exp_dir/numpy/splits_labels.npy \
$exp_dir/numpy \
- $exp_dir/data/${runDATA}.json
+ $exp_dir/data/${expname}.json
fi
## PCA SPACE AND PLOTS
@@ -86,23 +88,26 @@ if [ $stage == 4 ] || [ $stage == 0 ]; then
--output $exp_dir/numpy/pca \
--dim 3
printf "\n... PLOT similarity in PCA space (Analysis) ...\n"
+ mkdir -p $exp_dir/plots/analysis
python3 local/plot_pca_similarities.py \
$exp_dir/numpy/similarity_matrix.npy \
$exp_dir/numpy/splits_labels.npy \
$exp_dir/numpy/pca/3_pca_A.npy \
$exp_dir/numpy/pca/3_pca_b.npy \
- $exp_dir/plots \
- $exp_dir/data/${runDATA}.json
+ $exp_dir/plots/analysis \
+ $exp_dir/data/${expname}.json
fi
if [ $stage == 5 ] || [ $stage == 0 ]; then
printf "\n... Classification with k-means in PCA space ...\n"
+ mkdir -p $exp_dir/plots/kmeans
+ nclust=3 # number of clusters?
python3 local/classification_kmeans.py \
$exp_dir/numpy/similarity_matrix.npy \
$exp_dir/numpy/splits_labels.npy \
$exp_dir/numpy/pca/3_pca_A.npy \
$exp_dir/numpy/pca/3_pca_b.npy \
- $exp_dir/plots \
- $exp_dir/data/${runDATA}.json \
- 2
+ $exp_dir/plots/kmeans \
+ $exp_dir/data/${expname}.json \
+ $nclust
fi
\ No newline at end of file