Skip to content

change llama/modeling.py to opt npu performence #8342

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
merged 7 commits into from
Apr 30, 2024
Merged

change llama/modeling.py to opt npu performence #8342

Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
5db1b03
change llama/modeling.py to opt npu performence
wanghuancoder Apr 28, 2024
48a1338
update
wanghuancoder Apr 28, 2024
11a52e0
update
wanghuancoder Apr 28, 2024
f2252e8
Update modeling.py
Galaxy1458 Apr 29, 2024
198e58e
Merge branch 'develop' of https://github.com/PaddlePaddle/PaddleNLP i…
wanghuancoder Apr 30, 2024
e557b5a
Merge branch 'develop' of https://github.com/Galaxy1458/PaddleNLP int…
wanghuancoder Apr 30, 2024
672f4aa
add judge
wanghuancoder Apr 30, 2024
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
16 changes: 11 additions & 5 deletions paddlenlp/transformers/llama/modeling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -97,6 +97,7 @@ def swiglu(x, y=None):
"LlamaPretrainingCriterion",
]

npu_is_casual = False

def _get_interleave(n):
def _get_interleave_power_of_2(n):
Expand Down Expand Up @@ -244,7 +245,7 @@ def scaled_dot_product_attention(
attention_mask is None,
True,
False,
False,
npu_is_casual,
)[0]
else:
attn_output = F.scaled_dot_product_attention(
Expand Down Expand Up @@ -1118,6 +1119,7 @@ def __init__(self, config, layerwise_recompute: bool = False):
self.layerwise_recompute = layerwise_recompute
self.recompute_granularity = config.recompute_granularity


def forward(
self,
hidden_states: paddle.Tensor,
Expand Down Expand Up @@ -1612,11 +1614,12 @@ def forward(
attention_mask, (batch_size, seq_length), cache_length, inputs_embeds.dtype
) # [bs, 1, seq_len, seq_len]
if self.config.use_flash_attention:
is_casual = is_casual_mask(attention_mask)
if get_env_device() != "npu":
is_casual = is_casual_mask(attention_mask)
if is_casual and alibi is None:
attention_mask = None
else:
npu_is_casual = is_casual
attention_mask = attention_mask.astype("bool")
hidden_states = inputs_embeds
# decoder layers
Expand Down Expand Up @@ -1722,9 +1725,12 @@ def forward(self, prediction_scores, masked_lm_labels):
_hcg = fleet.get_hybrid_communicate_group()
masked_lm_loss = ConcatSePMaskedLoss.apply(masked_lm_loss, axis=1, group=_hcg.get_sep_parallel_group())
# skip ignore_index which loss == 0
masked_lm_loss = masked_lm_loss[masked_lm_loss > 0]
loss = paddle.mean(masked_lm_loss)

# masked_lm_loss = masked_lm_loss[masked_lm_loss > 0]
# loss = paddle.mean(masked_lm_loss)
binary_sequence = paddle.where(masked_lm_loss > 0, paddle.ones_like(masked_lm_loss), paddle.zeros_like(masked_lm_loss))
sum_ = paddle.sum(binary_sequence)
loss = 0 if sum_ == 0 else paddle.sum(masked_lm_loss * binary_sequence) / sum_

return loss


Expand Down