print("Training epoch: %d" % epoch)
print(eva.f1_score(labels_test_map, labels_predict_map))
report = eva.classification_report(labels_test_map, labels_predict_map)
print(report)
print(loss)
print(labels_test_map)
print(labels_predict_map)
print(transitions)
f_out.write("Training epoch: " + str(epoch) + "\n")
f_out.write("performance on entity extraction:" + "\n")
f_out.write("precision: " +
str(eva.precision_score(labels_test_map, labels_predict_map)))
f_out.write("\t" + "recall: " +
str(eva.recall_score(labels_test_map, labels_predict_map)))
f_out.write("\t" + "f1: " +
str(eva.f1_score(labels_test_map, labels_predict_map)))
f_out.write("\n" + report)
f_out.write("\n")
params = {
'model': model_trans.state_dict(),
'optim': optimizer_crf.state_dict()
}
torch.save(
params,
'/Users/MissyLinlin/Desktop/FYP/py/deepmaxsat/model/pretrain_trans_new.pt')
f_out.close()
posind_test, 1), model, entity=True, rel=True)
labels_test_map = [label_map[item] for sub in labels_test for item in sub]
labels_predict_map = [label_map[t.item()] for sub in labels_predict for t in sub]
print epoch
print eva.f1_score(labels_test_map, labels_predict_map)
report = eva.classification_report(labels_test_map, labels_predict_map)
p, r, f1, report_rel = eva.report_relation_trec(labels_rel_test, labels_rel_predict, 6)
print report
print report_rel
print loss
f_out.write("epoch: " + str(epoch) + "\n")
f_out.write("performance on entity extraction:" + "\n")
f_out.write("precision: " + str(eva.precision_score(labels_test_map, labels_predict_map)))
f_out.write("\t" + "recall: " + str(eva.recall_score(labels_test_map, labels_predict_map)))
f_out.write("\t" + "f1: " + str(eva.f1_score(labels_test_map, labels_predict_map)))
f_out.write("\n" + report)
f_out.write("\n")
f_out.write("performance on relation extraction:" + "\n")
f_out.write("precision: " + str(p))
f_out.write("\t" + "recall: " + str(r))
f_out.write("\t" + "f1: " + str(f1))
f_out.write('\n' + report_rel)
f_out.write("\n")
f_out.close()