explosion/spaCy
1
1
Fork 0
mirror of https://github.com/explosion/spaCy.git synced 2025-07-14 10:12:22 +03:00
Code Issues Packages Projects Releases Wiki Activity

chane naming and add finally block

Browse Source
This commit is contained in:
thomashacker 2023-03-27 13:58:36 +02:00
parent 8d50a8009c
commit 853ef78198
1 changed files with 19 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
spacy/training/pretrain.py
View File

@ -60,11 +60,11 @@ def pretrain(
row_settings = {"widths": (3, 10, 10, 6, 4), "aligns": ("r", "r", "r", "r", "r")} row_settings = {"widths": (3, 10, 10, 6, 4), "aligns": ("r", "r", "r", "r", "r")}
msg.row(("#", "# Words", "Total Loss", "Loss", "w/s"), **row_settings) msg.row(("#", "# Words", "Total Loss", "Loss", "w/s"), **row_settings)
def _save_model(epoch, is_temp=False, is_latest=False): def _save_model(epoch, is_temp=False, is_last=False):
is_temp_str = ".temp" if is_temp else "" is_temp_str = ".temp" if is_temp else ""
with model.use_params(optimizer.averages): with model.use_params(optimizer.averages):
if is_latest: if is_last:
save_path = output_dir / f"model_latest.bin" save_path = output_dir / f"model_last.bin"
else: else:
save_path = output_dir / f"model{epoch}{is_temp_str}.bin" save_path = output_dir / f"model{epoch}{is_temp_str}.bin"
with (save_path).open("wb") as file_: with (save_path).open("wb") as file_:
@ -81,6 +81,7 @@ def pretrain(
# TODO: I think we probably want this to look more like the # TODO: I think we probably want this to look more like the
# 'create_train_batches' function? # 'create_train_batches' function?
for epoch in range(epoch_resume, P["max_epochs"]): for epoch in range(epoch_resume, P["max_epochs"]):
try:
for batch_id, batch in enumerate(batcher(corpus(nlp))): for batch_id, batch in enumerate(batcher(corpus(nlp))):
docs = ensure_docs(batch) docs = ensure_docs(batch)
loss = make_update(model, docs, optimizer, objective) loss = make_update(model, docs, optimizer, objective)
@ -90,14 +91,14 @@ def pretrain(
if P["n_save_every"] and (batch_id % P["n_save_every"] == 0): if P["n_save_every"] and (batch_id % P["n_save_every"] == 0):
_save_model(epoch, is_temp=True) _save_model(epoch, is_temp=True)
if epoch + 1 == P["max_epochs"]: if P["n_save_epoch"]:
_save_model(epoch, is_latest=True)
elif P["n_save_epoch"]:
if epoch % P["n_save_epoch"] == 0 or epoch == P["max_epochs"] - 1: if epoch % P["n_save_epoch"] == 0 or epoch == P["max_epochs"] - 1:
_save_model(epoch) _save_model(epoch)
else: else:
_save_model(epoch) _save_model(epoch)
tracker.epoch_loss = 0.0 tracker.epoch_loss = 0.0
finally:
_save_model(epoch, is_last=True)
def ensure_docs(examples_or_docs: Iterable[Union[Doc, Example]]) -> List[Doc]: def ensure_docs(examples_or_docs: Iterable[Union[Doc, Example]]) -> List[Doc]: