Add option to SequenceGenerator to retain dropout (dccf7909) · Commits · Simon Will / fairseq

fairseq/models/fairseq_incremental_decoder.py

+1 −0

Changes for fairseq/models/fairseq_incremental_decoder.py: 1 added line, 0 removed lines.
Original line number	Diff line number	Diff line
		@@ -86,6 +86,7 @@ class FairseqIncrementalDecoder(FairseqDecoder):
		beam_size is required if using BeamableMM.
		"""
		if self._is_incremental_eval:
		del self._incremental_state
		self._incremental_state = {}

		def apply_clear_incremental_state(module):

+10 −4

Changes for fairseq/sequence_generator.py: 10 added lines, 4 removed lines.
Original line number	Diff line number	Diff line
		@@ -19,7 +19,7 @@ from fairseq.models import FairseqIncrementalDecoder
		class SequenceGenerator(object):
		def __init__(self, models, beam_size=1, minlen=1, maxlen=200,
		stop_early=True, normalize_scores=True, len_penalty=1,
		unk_penalty=0):
		unk_penalty=0, retain_dropout=False):
		"""Generates translations of a given source sentence.

		Args:
		@@ -45,6 +45,7 @@ class SequenceGenerator(object):
		self.normalize_scores = normalize_scores
		self.len_penalty = len_penalty
		self.unk_penalty = unk_penalty
		self.retain_dropout = retain_dropout

		def cuda(self):
		for model in self.models:
		@@ -98,6 +99,7 @@ class SequenceGenerator(object):

		encoder_outs = []
		for model in self.models:
		if not self.retain_dropout:
		model.eval()
		if isinstance(model.decoder, FairseqIncrementalDecoder):
		model.decoder.set_beam_size(beam_size)
		@@ -327,15 +329,19 @@ class SequenceGenerator(object):
		for model, encoder_out in zip(self.models, encoder_outs):
		decoder_out, attn = model.decoder(tokens, encoder_out)
		probs = F.softmax(decoder_out[:, -1, :], dim=1).data
		attn = attn[:, -1, :].data
		if avg_probs is None or avg_attn is None:
		if avg_probs is None:
		avg_probs = probs
		avg_attn = attn
		else:
		avg_probs.add_(probs)
		if attn is not None:
		attn = attn[:, -1, :].data
		if avg_attn is None:
		avg_attn = attn
		else:
		avg_attn.add_(attn)
		avg_probs.div_(len(self.models))
		avg_probs.log_()
		if avg_attn is not None:
		avg_attn.div_(len(self.models))

		return avg_probs, avg_attn