Skip to content

Support NNCF 2.14 and OV 2024.5 #997

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
Nov 21, 2024
Merged

Support NNCF 2.14 and OV 2024.5 #997

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
d2dc1c3
Tweak reference values
nikita-savelyevv Nov 12, 2024
6b8dd05
Update requirements to OV 2024.5
nikita-savelyevv Nov 12, 2024
5a70920
Update tokenizers condition in tests
nikita-savelyevv Nov 13, 2024
453f475
Fix condition
nikita-savelyevv Nov 14, 2024
588b03f
Merge branch 'main' into nncf-214-support
nikita-savelyevv Nov 15, 2024
2e0f0a2
Trigger Tests
nikita-savelyevv Nov 20, 2024
37c86d5
Trigger Tests
nikita-savelyevv Nov 21, 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
2 changes: 1 addition & 1 deletion setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,7 @@

EXTRAS_REQUIRE = {
"nncf": ["nncf>=2.11.0"],
"openvino": ["nncf>=2.11.0", "openvino==2024.4.1.dev20240926", "openvino-tokenizers==2024.4.1.0.dev20240926"],
"openvino": ["nncf>=2.11.0", "openvino==2024.5.0", "openvino-tokenizers==2024.5.0"],
"neural-compressor": ["neural-compressor[pt]>3.0", "accelerate", "transformers<4.46"],
"ipex": ["intel-extension-for-pytorch", "transformers>=4.39,<4.45"],
"diffusers": ["diffusers"],
Expand Down
15 changes: 8 additions & 7 deletions tests/openvino/test_exporters_cli.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,6 +50,7 @@
from optimum.intel.utils.import_utils import (
compare_versions,
is_openvino_tokenizers_available,
is_openvino_version,
is_tokenizers_version,
is_transformers_version,
)
Expand Down Expand Up @@ -80,19 +81,19 @@ class OVCLIExportTestCase(unittest.TestCase):
if is_transformers_version(">=", "4.45"):
SUPPORTED_ARCHITECTURES.extend([("text-to-image", "stable-diffusion-3"), ("text-to-image", "flux")])
EXPECTED_NUMBER_OF_TOKENIZER_MODELS = {
"gpt2": 2 if is_tokenizers_version("<", "0.20") else 0,
"gpt2": 2 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
"t5": 0, # no .model file in the repository
"albert": 0, # not supported yet
"distilbert": 1, # no detokenizer
"roberta": 2 if is_tokenizers_version("<", "0.20") else 0,
"roberta": 2 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
"vit": 0, # no tokenizer for image model
"wav2vec2": 0, # no tokenizer
"bert": 1, # no detokenizer
"blenderbot": 2 if is_tokenizers_version("<", "0.20") else 0,
"stable-diffusion": 2 if is_tokenizers_version("<", "0.20") else 0,
"stable-diffusion-xl": 4 if is_tokenizers_version("<", "0.20") else 0,
"stable-diffusion-3": 6 if is_tokenizers_version("<", "0.20") else 2,
"flux": 4 if is_tokenizers_version("<", "0.20") else 0,
"blenderbot": 2 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
"stable-diffusion": 2 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
"stable-diffusion-xl": 4 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
"stable-diffusion-3": 6 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 2,
"flux": 4 if is_tokenizers_version("<", "0.20") and is_openvino_version("<", "2024.5") else 0,
}

SUPPORTED_SD_HYBRID_ARCHITECTURES = [
Expand Down
4 changes: 2 additions & 2 deletions tests/openvino/test_quantization.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -251,7 +251,7 @@ class OVWeightCompressionTest(unittest.TestCase):
sensitivity_metric="mean_activation_magnitude",
dataset=["one two, " * i for i in range(10)],
),
{"int4": 25, "int8": 14},
{"int4": 24, "int8": 16},
),
(
OVModelForCausalLM,
Expand Down Expand Up @@ -876,7 +876,7 @@ def preprocess_function(examples, tokenizer):


class OVTrainerTest(unittest.TestCase):
SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS = (("albert", 64, 39),)
SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS = (("albert", 63, 39),)

@parameterized.expand(SUPPORTED_ARCHITECTURES_WITH_EXPECTED_QUANTIZED_MATMULS)
@unittest.skipIf(
Expand Down
Loading