explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-01-28 18:24:08 +03:00
Code Issues Packages Projects Releases Wiki Activity

small fixes

Browse Source
This commit is contained in:
svlandeg 2019-06-12 22:05:53 +02:00
parent 6521cfa132
commit b12001f368
2 changed files with 24 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
examples/pipeline/wiki_entity_linking/wiki_nel_pipeline.py
View File

@ -116,8 +116,8 @@ def run_pipeline():
# STEP 6: create the entity linking pipe # STEP 6: create the entity linking pipe
if train_pipe: if train_pipe:
print("STEP 6: training Entity Linking pipe", datetime.datetime.now()) print("STEP 6: training Entity Linking pipe", datetime.datetime.now())
train_limit = 100 train_limit = 5000
dev_limit = 20 dev_limit = 1000
print("Training on", train_limit, "articles") print("Training on", train_limit, "articles")
print("Dev testing on", dev_limit, "articles") print("Dev testing on", dev_limit, "articles")
print() print()
@ -145,6 +145,7 @@ def run_pipeline():
random.shuffle(train_data) random.shuffle(train_data)
losses = {} losses = {}
batches = minibatch(train_data, size=compounding(4.0, 128.0, 1.001)) batches = minibatch(train_data, size=compounding(4.0, 128.0, 1.001))
batchnr = 0
with nlp.disable_pipes(*other_pipes): with nlp.disable_pipes(*other_pipes):
for batch in batches: for batch in batches:
@ -156,35 +157,43 @@ def run_pipeline():
drop=DROPOUT, drop=DROPOUT,
losses=losses, losses=losses,
) )
batchnr += 1
except Exception as e: except Exception as e:
print("Error updating batch", e) print("Error updating batch", e)
losses['entity_linker'] = losses['entity_linker'] / batchnr
print("Epoch, train loss", itn, round(losses['entity_linker'], 2)) print("Epoch, train loss", itn, round(losses['entity_linker'], 2))
# baseline using only prior probabilities print()
el_pipe.context_weight = 0 print("STEP 7: performance measurement of Entity Linking pipe", datetime.datetime.now())
el_pipe.prior_weight = 1 print()
dev_acc_0_1 = _measure_accuracy(dev_data, el_pipe)
train_acc_0_1 = _measure_accuracy(train_data, el_pipe)
# print(" measuring accuracy 1-1") # print(" measuring accuracy 1-1")
el_pipe.context_weight = 1 el_pipe.context_weight = 1
el_pipe.prior_weight = 1 el_pipe.prior_weight = 1
dev_acc_1_1 = _measure_accuracy(dev_data, el_pipe) dev_acc_1_1 = _measure_accuracy(dev_data, el_pipe)
train_acc_1_1 = _measure_accuracy(train_data, el_pipe) train_acc_1_1 = _measure_accuracy(train_data, el_pipe)
print("train/dev acc combo:", round(train_acc_1_1, 2), round(dev_acc_1_1, 2))
# print(" measuring accuracy 1-0") # baseline using only prior probabilities
el_pipe.context_weight = 0
el_pipe.prior_weight = 1
dev_acc_0_1 = _measure_accuracy(dev_data, el_pipe)
train_acc_0_1 = _measure_accuracy(train_data, el_pipe)
print("train/dev acc prior:", round(train_acc_0_1, 2), round(dev_acc_0_1, 2))
# using only context
el_pipe.context_weight = 1 el_pipe.context_weight = 1
el_pipe.prior_weight = 0 el_pipe.prior_weight = 0
dev_acc_1_0 = _measure_accuracy(dev_data, el_pipe) dev_acc_1_0 = _measure_accuracy(dev_data, el_pipe)
train_acc_1_0 = _measure_accuracy(train_data, el_pipe) train_acc_1_0 = _measure_accuracy(train_data, el_pipe)
print("train/dev acc, 1-1, 0-1, 1-0:" , print("train/dev acc context:", round(train_acc_1_0, 2), round(dev_acc_1_0, 2))
round(train_acc_1_1, 2), round(train_acc_0_1, 2), round(train_acc_1_0, 2), "/", print()
round(dev_acc_1_1, 2), round(dev_acc_0_1, 2), round(dev_acc_1_0, 2))
# test Entity Linker
if to_test_pipeline: if to_test_pipeline:
print()
print("STEP 8: applying Entity Linking to toy example", datetime.datetime.now())
print() print()
run_el_toy_example(kb=my_kb, nlp=nlp) run_el_toy_example(kb=my_kb, nlp=nlp)
print() print()
@ -197,9 +206,9 @@ def _measure_accuracy(data, el_pipe):
correct = 0 correct = 0
incorrect = 0 incorrect = 0
docs = [d for d, g in data] docs = [d for d, g in data if len(d) > 0]
docs = el_pipe.pipe(docs) docs = el_pipe.pipe(docs)
golds = [g for d, g in data] golds = [g for d, g in data if len(d) > 0]
for doc, gold in zip(docs, golds): for doc, gold in zip(docs, golds):
try: try:

2
spacy/pipeline/pipes.pyx
View File

@ -1188,7 +1188,7 @@ class EntityLinker(Pipe):
def get_loss(self, docs, golds, scores): def get_loss(self, docs, golds, scores):
targets = [[1] for _ in golds] # assuming we're only using positive examples targets = [[1] for _ in golds] # assuming we're only using positive examples
loss, gradients = self.get_cossim_loss_2(yh=scores, y=golds, t=targets) loss, gradients = self.get_cossim_loss_2(yh=scores, y=golds, t=targets)
#loss = loss / len(golds) loss = loss / len(golds)
return loss, gradients return loss, gradients
def get_cossim_loss_2(self, yh, y, t): def get_cossim_loss_2(self, yh, y, t):