def load_model(path_to_model,path_to_tables):
"""
Load the model with saved tables
"""
# Load model options
print ('Loading model parameters...')
with open('%s.pkl'%path_to_model, 'rb') as f:
options = pkl.load(f)
# Load parameters
params = init_params(options)
params = load_params(path_to_model, params)
tparams = init_tparams(params)
# Extractor functions
print ('Compiling encoders...')
embedding, x_mask, ctxw2v = build_encoder(tparams, options)
f_w2v = theano.function([embedding, x_mask], ctxw2v, name='f_w2v')
# Tables
print ('Loading tables...')
table = load_tables()
# Store everything we need in a dictionary
print ('Packing up...')
model = {}
model['options'] = options
model['table'] = table
model['f_w2v'] = f_w2v
return model
def load_model(embed_map=None):
"""
Load all model components + apply vocab expansion
"""
# Load the worddict
print('Loading dictionary...')
with open(config.paths['dictionary'], 'rb') as f:
worddict = pkl.load(f)
# Create inverted dictionary
print('Creating inverted dictionary...')
word_idict = dict()
for kk, vv in worddict.items():
word_idict[vv] = kk
word_idict[0] = '<eos>'
word_idict[1] = 'UNK'
# Load model options
print('Loading model options...')
with open('%s.pkl' % config.paths['skmodels'], 'rb') as f:
options = pkl.load(f)
# Load parameters
print('Loading model parameters...')
params = init_params(options)
params = load_params(config.paths['skmodels'], params)
tparams = init_tparams(params)
# Extractor functions
print('Compiling encoder...')
trng = RandomStreams(1234)
trng, x, x_mask, ctx, emb = build_encoder(tparams, options)
f_enc = theano.function([x, x_mask], ctx, name='f_enc')
f_emb = theano.function([x], emb, name='f_emb')
trng, embedding, x_mask, ctxw2v = build_encoder_w2v(tparams, options)
f_w2v = theano.function([embedding, x_mask], ctxw2v, name='f_w2v')
# Load word2vec, if applicable
if embed_map == None:
print('Loading word2vec embeddings...')
embed_map = load_googlenews_vectors(config.paths['v_expansion'])
# Lookup table using vocab expansion trick
print('Creating word lookup tables...')
table = lookup_table(options, embed_map, worddict, word_idict, f_emb)
# Store everything we need in a dictionary
print('Packing up...')
model = {}
model['options'] = options
model['table'] = table
model['f_w2v'] = f_w2v
return model
def load_model():
"""
Load a trained model for decoding
"""
# Load the worddict
print('Loading dictionary...')
with open(config.paths['dictionary'], 'rb') as f:
worddict = pkl.load(f)
# Create inverted dictionary
print('Creating inverted dictionary...')
word_idict = dict()
for kk, vv in worddict.items():
word_idict[vv] = kk
word_idict[0] = '<eos>'
word_idict[1] = 'UNK'
# Load model options
print('Loading model options...')
with open('%s.pkl' % config.paths['skvmodels'], 'rb') as f:
options = pkl.load(f)
# Load parameters
print('Loading model parameters...')
params = init_params(options)
params = load_params(config.paths['skvmodels'], params)
tparams = init_tparams(params)
# Sampler.
trng = RandomStreams(1234)
f_init, f_next = build_sampler(tparams, options, trng)
# Pack everything up
dec = dict()
dec['options'] = options
dec['trng'] = trng
dec['worddict'] = worddict
dec['word_idict'] = word_idict
dec['tparams'] = tparams
dec['f_init'] = f_init
dec['f_next'] = f_next
return dec
def load_model(path_to_model, path_to_dictionary):
"""
Load a trained model for decoding
"""
# Load the worddict
with open(path_to_dictionary, 'rb') as f:
worddict = pkl.load(f)
# Create inverted dictionary
word_idict = dict()
for kk, vv in worddict.items():
word_idict[vv] = kk
word_idict[0] = '<eos>'
word_idict[1] = 'UNK'
# Load model options
with open('%s.pkl'%path_to_model, 'rb') as f:
options = pkl.load(f)
if 'doutput' not in options.keys():
options['doutput'] = True
# Load parameters
params = init_params(options)
params = load_params(path_to_model, params)
tparams = init_tparams(params)
# Sampler.
trng = RandomStreams(1234)
f_init, f_next = build_sampler(tparams, options, trng)
# Pack everything up
dec = dict()
dec['options'] = options
dec['trng'] = trng
dec['worddict'] = worddict
dec['word_idict'] = word_idict
dec['tparams'] = tparams
dec['f_init'] = f_init
dec['f_next'] = f_next
return dec
def load_model(path_to_model):
"""
Load all model components
"""
# Load the worddict
with open('%s.dictionary.pkl' % path_to_model, 'rb') as f:
worddict = pkl.load(f)
# Create inverted dictionary
word_idict = dict()
for kk, vv in worddict.items():
word_idict[vv] = kk
word_idict[0] = '<eos>'
word_idict[1] = 'UNK'
# Load model options
with open('%s.pkl' % path_to_model, 'rb') as f:
options = pkl.load(f)
# Load parameters
params = init_params(options)
params = load_params(path_to_model, params)
tparams = init_tparams(params)
# Extractor functions
trng = RandomStreams(1234)
trng, [x, x_mask], sentences = build_sentence_encoder(tparams, options)
f_senc = theano.function([x, x_mask], sentences, name='f_senc')
trng, [im], images = build_image_encoder(tparams, options)
f_ienc = theano.function([im], images, name='f_ienc')
# Store everything we need in a dictionary
model = {}
model['options'] = options
model['worddict'] = worddict
model['word_idict'] = word_idict
model['f_senc'] = f_senc
model['f_ienc'] = f_ienc
return model