def evalu(rdnn, ddat, dset):
cost, preds = rdnn.predict(ddat)
preds = [p.argmax(axis=-1).flatten() for b in preds for p in b]
correct = sum(np.sum(np.array(sent['lseq']) == pred) for sent, pred in zip(dset,preds))
total = sum(len(p) for p in preds)
cerr = 1-(correct/total)
gold_labels = []
pred_labels = []
for sent, pred in zip(dset,preds):
ls = get_ls(sent['wiseq'],pred)
gold_labels.extend([sent['ls'][ii] for ii, interested in enumerate(sent['ii']) if interested])
pred_labels.extend([ls[ii] for ii, interested in enumerate(sent['ii']) if interested])
p, r, f = evaluate_system.evaluateIdentifier(gold_labels, pred_labels)
return [cost, cerr, p, r, f]
def evalu(dset, yhat):
# slens = [len(sent['ls']) for sent in dset]
slens = [sum(sent['ii']) for sent in dset]
offset = 0
preds = []
for slen in slens:
preds.append(yhat[offset:offset+slen])
offset+=slen
assert len(preds) == len(dset)
gold_labels = []
pred_labels = []
for sent, pred in zip(dset,preds):
gold_labels.extend([sent['ls'][ii] for ii, interested in enumerate(sent['ii']) if interested])
# pred_labels.extend([pred[ii] for ii, interested in enumerate(sent['ii']) if interested])
pred_labels.extend(pred)
logging.debug(tabulate(confusion_matrix(np.array(gold_labels), np.array(pred_labels)), headers=[0,1]))
p, r, f = evaluate_system.evaluateIdentifier(gold_labels, pred_labels)
return p,r,f
