riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-31 02:02:21 +06:00
Code Issues Actions 23 Packages Projects Releases Wiki Activity

style

Browse Source
This commit is contained in:
Cyril Vallez 2025-07-03 16:06:19 +02:00
parent 810d33f40b
commit 554be7dac3
No known key found for this signature in database
1 changed files with 13 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
tests/utils/test_masking_utils.py
View File

@ -16,12 +16,15 @@ import unittest
from transformers.testing_utils import is_torch_available, require_torch
if is_torch_available():
import torch
from transformers import AutoConfig
from transformers.masking_utils import create_causal_mask, create_sliding_window_causal_mask
from transformers import AutoConfig
from transformers.masking_utils import create_causal_mask
# fmt: off
EXPECTED_PACKED_MASK = torch.tensor([[[
[ True, False, False, False, False, False, False, False, False, False],
[ True, True, False, False, False, False, False, False, False, False],
@ -35,7 +38,7 @@ EXPECTED_PACKED_MASK = torch.tensor([[[
[False, False, False, False, False, False, True, True, True, True]]],
[[[ True, False, False, False, False, False, False, False, False, False],
[[[ True, False, False, False, False, False, False, False, False, False],
[ True, True, False, False, False, False, False, False, False, False],
[ True, True, True, False, False, False, False, False, False, False],
[ True, True, True, True, False, False, False, False, False, False],
@ -46,13 +49,12 @@ EXPECTED_PACKED_MASK = torch.tensor([[[
[False, False, False, False, False, False, True, True, True, False],
[False, False, False, False, False, False, True, True, True, True]
]]], dtype=torch.bool)
# fmt: on
@require_torch
class MaskTest(unittest.TestCase):
def test_packed_sequence_mask_sdpa(self):
config = AutoConfig.from_pretrained("meta-llama/Llama-3.2-1B")
config._attn_implementation = "sdpa"
@ -61,7 +63,7 @@ class MaskTest(unittest.TestCase):
cache_position = torch.arange(sequence_length)
# First batch has 3 packed sequences of 4, 2 and 4 tokens respectively, second has 2 of 6 and 4 tokens
position_ids = torch.tensor([[0,1,2,3,0,1,0,1,2,3], [0,1,2,3,4,5,0,1,2,3,]])
position_ids = torch.tensor([[0, 1, 2, 3, 0, 1, 0, 1, 2, 3], [0, 1, 2, 3, 4, 5, 0, 1, 2, 3]])
causal_mask = create_causal_mask(
config=config,
@ -76,7 +78,6 @@ class MaskTest(unittest.TestCase):
self.assertEqual(causal_mask, EXPECTED_PACKED_MASK)
def test_packed_sequence_mask_eager(self):
config = AutoConfig.from_pretrained("meta-llama/Llama-3.2-1B")
config._attn_implementation = "eager"
@ -85,7 +86,7 @@ class MaskTest(unittest.TestCase):
cache_position = torch.arange(sequence_length)
# First batch has 3 packed sequences of 4, 2 and 4 tokens respectively, second has 2 of 6 and 4 tokens
position_ids = torch.tensor([[0,1,2,3,0,1,0,1,2,3], [0,1,2,3,4,5,0,1,2,3,]])
position_ids = torch.tensor([[0, 1, 2, 3, 0, 1, 0, 1, 2, 3], [0, 1, 2, 3, 4, 5, 0, 1, 2, 3]])
causal_mask = create_causal_mask(
config=config,
@ -98,10 +99,9 @@ class MaskTest(unittest.TestCase):
)
min_dtype = torch.finfo(torch.float16).min
self.assertEqual(causal_mask, torch.where(EXPECTED_PACKED_MASK, 0., min_dtype))
self.assertEqual(causal_mask, torch.where(EXPECTED_PACKED_MASK, 0.0, min_dtype))
def test_packed_sequence_mask_flex_attention(self):
config = AutoConfig.from_pretrained("meta-llama/Llama-3.2-1B")
config._attn_implementation = "flex_attention"
@ -110,7 +110,7 @@ class MaskTest(unittest.TestCase):
cache_position = torch.arange(sequence_length)
# First batch has 3 packed sequences of 4, 2 and 4 tokens respectively, second has 2 of 6 and 4 tokens
position_ids = torch.tensor([[0,1,2,3,0,1,0,1,2,3], [0,1,2,3,4,5,0,1,2,3,]])
position_ids = torch.tensor([[0, 1, 2, 3, 0, 1, 0, 1, 2, 3], [0, 1, 2, 3, 4, 5, 0, 1, 2, 3]])
causal_mask = create_causal_mask(
config=config,
@ -123,4 +123,4 @@ class MaskTest(unittest.TestCase):
)
block_mask = causal_mask.to_dense()
self.assertEqual(block_mask, EXPECTED_PACKED_MASK)
self.assertEqual(block_mask, EXPECTED_PACKED_MASK)