Skip to content

Update FB15k_evaluation.py #1

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
while519 merged 1 commit into from
May 17, 2016
Merged

Update FB15k_evaluation.py #1

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 27 additions & 2 deletions FB15k/FB15k_evaluation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
#! /usr/bin/python
import sys
sys.path.append('/volatile/yuwu/git-repo/SME')
from model import *

def load_file(path):
Expand Down Expand Up @@ -30,12 +31,31 @@ def RankingEval(datapath='../data/', dataset='FB15k-test',
o = load_file(datapath + dataset + '-rel.pkl')
if type(embeddings) is list:
o = o[-embeddings[1].N:, :]

tl = load_file(datapath + dataset + '-train-lhs.pkl')
tr = load_file(datapath + dataset + '-train-rhs.pkl')
to = load_file(datapath + dataset + '-train-rel.pkl')
if type(embeddings) is list:
to = to[-embeddings[1].N:, :]


vl = load_file(datapath + dataset + '-valid-lhs.pkl')
vr = load_file(datapath + dataset + '-valid-rhs.pkl')
vo = load_file(datapath + dataset + '-valid-rel.pkl')
if type(embeddings) is list:
vo = vo[-embeddings[1].N:, :]

# Convert sparse matrix to indexes
if neval == 'all':
idxl = convert2idx(l)
idxr = convert2idx(r)
idxo = convert2idx(o)
idxtl = convert2idx(tl)
idxtr = convert2idx(tr)
idxto = convert2idx(to)
idxvl = convert2idx(vl)
idxvr = convert2idx(vr)
idxvo = convert2idx(vo)
else:
idxl = convert2idx(l)[:neval]
idxr = convert2idx(r)[:neval]
Expand All @@ -45,8 +65,13 @@ def RankingEval(datapath='../data/', dataset='FB15k-test',
subtensorspec=Nsyn)
rankrfunc = RankRightFnIdx(simfn, embeddings, leftop, rightop,
subtensorspec=Nsyn)

true_triples=np.concatenate([idxtl,idxvl,idxl,idxto,idxvo,idxo,idxtr,idxvr,idxr]).reshape(3,idxtl.shape[0]+idxvl.shape[0]+idxl.shape[0]).T

res = RankingScoreIdx(ranklfunc, rankrfunc, idxl, idxr, idxo)
res = FilteredRankingScoreIdx(ranklfunc, rankrfunc, idxl, idxr, idxo, true_triples)
f = open('Filtered_eval.pkl', 'wb')
cPickle.dump(res, f, -1)
f.close()
dres = {}
dres.update({'microlmean': np.mean(res[0])})
dres.update({'microlmedian': np.median(res[0])})
Expand Down Expand Up @@ -208,4 +233,4 @@ def RankingEvalFil(datapath='../data/', dataset='umls-test', op='TransE',
return MR, T10

if __name__ == '__main__':
RankingEval(loadmodel=sys.argv[1])
RankingEval(loadmodel='FB15k_TransE/best_valid_model.pkl')