Skip to content
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

Skip BloomEmbeddingTest.test_embeddings for PT < 1.10 #19261

Merged
merged 1 commit into from
Oct 10, 2022
Merged

Skip BloomEmbeddingTest.test_embeddings for PT < 1.10 #19261

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
8 changes: 7 additions & 1 deletion tests/models/bloom/test_modeling_bloom.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@
BloomModel,
BloomTokenizerFast,
)
from transformers.pytorch_utils import is_torch_greater_or_equal_than_1_10


@require_torch
Expand Down Expand Up @@ -490,9 +491,14 @@ def setUp(self):
super().setUp()
self.path_bigscience_model = "bigscience/bigscience-small-testing"

@unittest.skipIf(
not is_torch_available() or not is_torch_greater_or_equal_than_1_10,
"Test failed with torch < 1.10 (`LayerNormKernelImpl` not implemented for `BFloat16`)",
)
@require_torch
def test_embeddings(self):
model = BloomForCausalLM.from_pretrained(self.path_bigscience_model, torch_dtype="auto") # load in fp32
# The config in this checkpoint has `bfloat16` as `torch_dtype` -> model in `bfloat16`
model = BloomForCausalLM.from_pretrained(self.path_bigscience_model, torch_dtype="auto")
model.eval()

EMBEDDINGS_DS_BEFORE_LN_BF_16_MEAN = {
Expand Down