mirror of
https://github.com/cmusphinx/sphinxtrain.git
synced 2026-05-17 13:10:52 +00:00
David Huggins-Daines
90 lines
2.9 KiB
Python
90 lines
2.9 KiB
Python
#!/usr/bin/env python
|
|
|
|
import sys
|
|
from cmusphinx import s3mixw
|
|
import struct
|
|
import numpy
|
|
|
|
|
|
class Sendump:
|
|
def __init__(self, filename):
|
|
self.integer_format = None
|
|
self.load(filename)
|
|
|
|
def unpack_endian(self, s):
|
|
if self.integer_format is None:
|
|
self.integer_format = 'I'
|
|
nbytes = struct.unpack(self.integer_format, s)[0]
|
|
if nbytes > 65536:
|
|
self.integer_format = '!I'
|
|
nbytes = struct.unpack(self.integer_format, s)[0]
|
|
return nbytes
|
|
|
|
def readstr(self, fh):
|
|
nbytes = self.unpack_endian(fh.read(4))
|
|
if nbytes == 0:
|
|
return None
|
|
else:
|
|
return fh.read(nbytes)[0:-1]
|
|
|
|
def mixw(self):
|
|
return self.opdf
|
|
|
|
def load(self, filename):
|
|
mixture_count = 0
|
|
model_count = 0
|
|
mixw_shift = 10
|
|
cluster_bits = 8
|
|
feature_count = 0
|
|
cluster_count = 0
|
|
|
|
sendump = open(filename, "rb")
|
|
_ = self.readstr(sendump) # title
|
|
while True:
|
|
header = self.readstr(sendump)
|
|
if header is None:
|
|
break
|
|
if header.startswith(b"mixture_count"):
|
|
mixture_count = (int)(header.split()[1])
|
|
if header.startswith(b"model_count"):
|
|
model_count = (int)(header.split()[1])
|
|
if header.startswith(b"mixw_shift"):
|
|
mixw_shift = (int)(header.split()[1])
|
|
if header.startswith(b"cluster_bits"):
|
|
cluster_bits = (int)(header.split()[1])
|
|
if header.startswith(b"feature_count"):
|
|
feature_count = (int)(header.split()[1])
|
|
if header.startswith(b"cluster_count"):
|
|
cluster_count = (int)(header.split()[1])
|
|
|
|
if cluster_count == 0:
|
|
# Number of codewords and pdfs
|
|
model_count = self.unpack_endian(sendump.read(4))
|
|
mixture_count = self.unpack_endian(sendump.read(4))
|
|
print("rows (model_count): %d, columns (mixture_count): %d, "
|
|
"features (feature_count): %d"
|
|
% (model_count, mixture_count, feature_count))
|
|
|
|
# Now read the stuff
|
|
self.opdf = numpy.empty((mixture_count, feature_count, model_count))
|
|
for i in range(0, feature_count):
|
|
for j in range(0, model_count):
|
|
# Read bytes, expand to ints, shift them up
|
|
mixw = numpy.fromfile(sendump, 'B',
|
|
mixture_count).astype('i') << mixw_shift
|
|
# Negate, exponentiate, and untranspose
|
|
self.opdf[:, i, j] = numpy.power(1.0001, -mixw)
|
|
|
|
|
|
def usage():
|
|
print("Usage: %s IN_SENDUMP OUT_MIXW" % sys.argv[0])
|
|
|
|
|
|
if __name__ == '__main__':
|
|
if len(sys.argv) < 3:
|
|
usage()
|
|
sys.exit(2)
|
|
sendump = Sendump(sys.argv[1])
|
|
with s3mixw.open(sys.argv[2], 'wb') as s3m:
|
|
s3m.writeall(sendump.mixw())
|