def load_model(trainf, model_name, weights_file):
params = []
module = importlib.import_module('.'+modelname, 'models')
conf, ps, h = anssel_train.config(module.config, params)
s0, s1, y, vocab, gr = anssel_train.load_set(trainf)
model = anssel_train.build_model(glove, vocab, module.prep_model, conf)
model.load_weights(weights_file)
return model, vocab
def load_model(trainf, model_name, weights_file):
params = []
module = importlib.import_module('.' + modelname, 'models')
conf, ps, h = anssel_train.config(module.config, params)
s0, s1, y, vocab, gr = anssel_train.load_set(trainf)
model = anssel_train.build_model(glove, vocab, module.prep_model, conf)
model.load_weights(weights_file)
return model, vocab
def get_score():
if (request.json['atext'] == []):
return jsonify({'score': []}), 200
f = tempfile.NamedTemporaryFile(mode='w')
# FIXME: Avoid temporary files!!!
write_csv(f.file, load_samples(request.json['qtext'], request.json['atext']))
f.file.close()
s0t, s1t, yt, _, grt = anssel_train.load_set(f.name, vocab, skip_oneclass=False)
res = model.predict(grt)['score'][:,0]
return jsonify({'score': res.tolist()}), 200
def get_score():
if (request.json['atext'] == []):
return jsonify({'score': []}), 200
f = tempfile.NamedTemporaryFile(mode='w')
# FIXME: Avoid temporary files!!!
write_csv(f.file, load_samples(request.json['qtext'],
request.json['atext']))
f.file.close()
s0t, s1t, yt, _, grt = anssel_train.load_set(f.name,
vocab,
skip_oneclass=False)
res = model.predict(grt)['score'][:, 0]
return jsonify({'score': res.tolist()}), 200
print('Predict&Eval (best epoch)')
model.load_weights('weights-'+runid+'-bestval.h5')
ev.eval_anssel(model.predict(grv)['score'][:,0], grv['si0'], grv['score'], 'anssel Val')
if __name__ == "__main__":
modelname, weightsf, vocabf, trainf, valf = sys.argv[1:6]
params = sys.argv[6:]
module = importlib.import_module('.'+modelname, 'models')
conf, ps, h = config(module.config, params)
runid = '%s-%x' % (modelname, h)
print('RunID: %s (%s)' % (runid, ps))
print('GloVe')
glove = emb.GloVe(N=conf['embdim'])
print('Dataset (vocab)')
vocab = pickle.load(open(vocabf, "rb")) # use plain pickle because unicode
print('Dataset (anssel train)')
s0, s1, y, _, gr_ = anssel_train.load_set(trainf, vocab, s0pad=s0pad, s1pad=s1pad)
gr = loader.graph_adapt_ubuntu(gr_, vocab)
print('Dataset (anssel val)')
s0v, s1v, yv, _, grv_ = anssel_train.load_set(valf, vocab, s0pad=s0pad, s1pad=s1pad)
grv = loader.graph_adapt_ubuntu(grv_, vocab)
transfer_eval(runid, weightsf, module.prep_model, conf, glove, vocab, gr, grv)
print('%d 0 %d 1 %f %s' % (n, m, iy, code), file=f)
m += 1
if __name__ == "__main__":
modelname, weightsfile, trainf, valf, trec_qrels_file, trec_top_file = sys.argv[1:7]
params = sys.argv[7:]
module = importlib.import_module('.'+modelname, 'models')
conf, ps, h = anssel_train.config(module.config, params)
print('GloVe')
glove = emb.GloVe(N=conf['embdim'])
print('Dataset')
s0, s1, y, vocab, gr = anssel_train.load_set(trainf)
s0t, s1t, yt, _, grt = anssel_train.load_set(valf, vocab)
print('Model')
model = anssel_train.build_model(glove, vocab, module.prep_model, conf)
print('Weights')
model.load_weights(weightsfile)
print('Prediction')
ypred = model.predict(gr)['score'][:,0]
ypredt = model.predict(grt)['score'][:,0]
ev.eval_anssel(ypred, s0, y, trainf)
ev.eval_anssel(ypredt, s0t, yt, valf)
from keras.layers.recurrent import SimpleRNN, GRU, LSTM
import pysts.embedding as emb
import pysts.eval as ev
import pysts.kerasts.blocks as B
from pysts.kerasts.callbacks import AnsSelCB
from pysts.kerasts.objectives import ranknet
from pysts.hyperparam import RandomSearch
import anssel_train
if __name__ == "__main__":
modelname, trainf, valf = sys.argv[1:4]
module = importlib.import_module('.' + modelname, 'models')
s0, s1, y, vocab, gr = anssel_train.load_set(trainf)
s0t, s1t, yt, _, grt = anssel_train.load_set(valf, vocab)
glove = emb.GloVe(300) # XXX hardcoded N
rs = RandomSearch(modelname + '_rlog.txt',
dropout=[1 / 2, 2 / 3, 3 / 4],
inp_e_dropout=[1 / 2, 3 / 4, 4 / 5],
l2reg=[1e-4, 1e-3, 1e-2],
cnnact=['tanh', 'tanh', 'relu'],
cnninit=['glorot_uniform', 'glorot_uniform', 'normal'],
cdim={
1: [0, 0, 1 / 2, 1, 2],
2: [0, 0, 1 / 2, 1, 2, 0],
3: [0, 0, 1 / 2, 1, 2, 0],
4: [0, 0, 1 / 2, 1, 2, 0],
if __name__ == "__main__":
niter, modelname, trainf, valf, testf = sys.argv[1:6]
params = sys.argv[6:]
niter = int(niter)
module = importlib.import_module('.'+modelname, 'models')
conf, ps, h = anssel_train.config(module.config, params)
if conf['embdim'] is not None:
print('GloVe')
glove = emb.GloVe(N=conf['embdim'])
else:
glove = None
print('Dataset')
s0, s1, y, vocab, gr = anssel_train.load_set(trainf)
s0v, s1v, yv, _, grv = anssel_train.load_set(valf, vocab)
s0t, s1t, yt, _, grt = anssel_train.load_set(testf, vocab)
mrr = []
mrrv = []
mrrt = []
mapt = []
for i in range(niter):
runid = '%s-%x-%02d' % (modelname, h, i)
print('RunID: %s (%s)' % (runid, ps))
model = anssel_train.train_and_eval(runid, module.prep_model, conf, glove, vocab, gr, s0, grv, s0v, do_eval=False)
print('Predict&Eval (best val epoch)')
ypred = model.predict(gr)['score'][:,0]