for i in range(num_chunks):
start = i * (args.chunk_size / 2)
this_chunk_size = args.chunk_size if feature.shape[
0] - start > args.chunk_size else feature.shape[0] - start
feature_length.append(this_chunk_size)
feature_array.append(feature[start:start + this_chunk_size])
feature_length = np.expand_dims(np.array(feature_length), axis=1)
# Except for the last feature, the length of other features should be the same (=chunk_size)
embeddings = trainer.predict(
np.array(feature_array[:-1], dtype=np.float32))
embedding_last = trainer.predict(feature_array[-1])
embeddings = np.concatenate(
[embeddings,
np.expand_dims(embedding_last, axis=0)], axis=0)
if args.normalize:
embeddings /= np.sqrt(
np.sum(np.square(embeddings), axis=1, keepdims=True))
embedding = np.sum(embeddings * feature_length,
axis=0) / np.sum(feature_length)
else:
tf.logging.info("[INFO] Key %s length %d." %
(key, feature.shape[0]))
embedding = trainer.predict(feature)
if args.normalize:
embedding /= np.sqrt(np.sum(np.square(embedding)))
write_vec_flt(fp_out, embedding, key=key)
fp_out.close()
trainer.close()
vector_in = sys.argv[1:-1]
vector_out = sys.argv[-1]
vector_tot = []
vector_names = []
vector_single = {}
for key, vec in read_vec_flt_scp(vector_in[0]):
vector_names.append(key)
vector_single[key] = vec
dim = vec.shape[0]
vector_tot.append(vector_single)
for vector_file in vector_in[1:]:
vector_single = {}
index = 0
for key, vec in read_vec_flt_scp(vector_file):
assert(key == vector_names[index])
index += 1
vector_single[key] = vec
dim = vec.shape[0]
vector_tot.append(vector_single)
with open_or_fd(vector_out, 'wb') as f:
for name in vector_names:
vector = []
for vector_single in vector_tot:
vector.append(vector_single[name])
vector = np.concatenate(vector, axis=0)
write_vec_flt(f, vector, key=name)