def reduce_metrics(logging_outputs) -> None:
"""Aggregate logging outputs from data parallel training."""
loss_sum = utils.item(sum(log.get("loss", 0) for log in logging_outputs))
nll_loss_sum = utils.item(
sum(log.get("nll_loss", 0) for log in logging_outputs)
)
alignment_loss_sum = utils.item(
sum(log.get("alignment_loss", 0) for log in logging_outputs)
)
ntokens = utils.item(sum(log.get("ntokens", 0) for log in logging_outputs))
sample_size = utils.item(
sum(log.get("sample_size", 0) for log in logging_outputs)
)
metrics.log_scalar(
"loss", loss_sum / sample_size / math.log(2), sample_size, round=3
)
metrics.log_scalar(
"nll_loss", nll_loss_sum / ntokens / math.log(2), ntokens, round=3
)
metrics.log_scalar(
"alignment_loss",
alignment_loss_sum / sample_size / math.log(2),
sample_size,
round=3,
)
metrics.log_derived(
"ppl", lambda meters: utils.get_perplexity(meters["nll_loss"].avg)
)
def reduce_metrics(logging_outputs) -> None:
"""Aggregate logging outputs from data parallel training."""
sample_size = utils.item(
sum(log.get("sample_size", 0) for log in logging_outputs))
loss = utils.item(sum(log.get("loss", 0) for log in logging_outputs))
nll_loss = utils.item(
sum(log.get("nll_loss", 0) for log in logging_outputs))
metrics.log_scalar("loss",
loss / sample_size / math.log(2),
sample_size,
round=3)
metrics.log_scalar("nll_loss",
nll_loss / sample_size / math.log(2),
sample_size,
round=3)
metrics.log_derived(
"ppl", lambda meters: utils.get_perplexity(meters["loss"].avg))
for key in logging_outputs[0]:
if key[-5:] == "-loss":
val = sum(log.get(key, 0) for log in logging_outputs)
metrics.log_scalar(
key[:-5],
val / sample_size /
math.log(2) if sample_size > 0 else 0.0,
sample_size,
round=3,
)
def reduce_metrics(cls, logging_outputs) -> None:
"""Aggregate logging outputs from data parallel training."""
loss_sum = sum(log.get("loss", 0) for log in logging_outputs)
nll_loss_sum = sum(log.get("nll_loss", 0) for log in logging_outputs)
ntokens = sum(log.get("ntokens", 0) for log in logging_outputs)
sample_size = sum(log.get("sample_size", 0) for log in logging_outputs)
metrics.log_scalar(
"loss", loss_sum / sample_size / math.log(2), sample_size, round=3
)
metrics.log_scalar(
"nll_loss", nll_loss_sum / ntokens / math.log(2), ntokens, round=3
)
metrics.log_derived(
"ppl", lambda meters: utils.get_perplexity(meters["nll_loss"].avg)
)
total = utils.item(sum(log.get("total", 0) for log in logging_outputs))
if total > 0:
metrics.log_scalar("total", total)
n_correct = utils.item(
sum(log.get("n_correct", 0) for log in logging_outputs)
)
metrics.log_scalar("n_correct", n_correct)
metrics.log_derived(
"accuracy",
lambda meters: round(
meters["n_correct"].sum * 100.0 / meters["total"].sum, 3
)
if meters["total"].sum > 0
else float("nan"),
)
def reduce_metrics(logging_outputs) -> None:
"""Aggregate logging outputs from data parallel training."""
loss_sum = sum(log.get("loss", 0) for log in logging_outputs)
sample_size = sum(log.get("sample_size", 0) for log in logging_outputs)
metrics.log_scalar("loss",
loss_sum / sample_size / math.log(2),
sample_size,
round=3)
metrics.log_derived(
"ppl", lambda meters: utils.get_perplexity(meters["loss"].avg))
def reduce_metrics(logging_outputs) -> None:
"""Aggregate logging outputs from data parallel training."""
loss_sum = sum(log.get("loss", 0) for log in logging_outputs)
ntokens = sum(log.get("ntokens", 0) for log in logging_outputs)
sample_size = sum(log.get("sample_size", 0) for log in logging_outputs)
# we divide by log(2) to convert the loss from base e to base 2
metrics.log_scalar("loss",
loss_sum / sample_size / math.log(2),
sample_size,
round=3)
if sample_size != ntokens:
metrics.log_scalar("nll_loss",
loss_sum / ntokens / math.log(2),
ntokens,
round=3)
metrics.log_derived(
"ppl",
lambda meters: utils.get_perplexity(meters["nll_loss"].avg))
else:
metrics.log_derived(
"ppl", lambda meters: utils.get_perplexity(meters["loss"].avg))