-
Notifications
You must be signed in to change notification settings - Fork 6
/
Copy pathtest_modeling_gemma2.py
49 lines (42 loc) · 1.98 KB
/
test_modeling_gemma2.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
# coding=utf-8
# Copyright 2024 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
import pytest
from executorch.extension.pybindings.portable_lib import ExecuTorchModule
from transformers import AutoTokenizer
from transformers.testing_utils import slow
from optimum.executorch import ExecuTorchModelForCausalLM
class ExecuTorchModelIntegrationTest(unittest.TestCase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
@slow
@pytest.mark.run_slow
def test_gemma2_text_generation_with_xnnpack(self):
# TODO: Switch to use google/gemma-2-2b once https://github.com/huggingface/optimum/issues/2127 is fixed
# model_id = "google/gemma-2-2b"
model_id = "unsloth/gemma-2-2b-it"
model = ExecuTorchModelForCausalLM.from_pretrained(model_id, recipe="xnnpack")
self.assertIsInstance(model, ExecuTorchModelForCausalLM)
self.assertIsInstance(model.model, ExecuTorchModule)
EXPECTED_GENERATED_TEXT = (
"Hello I am doing a project for my school and I need to make sure it is a great to be creative and I can!"
)
tokenizer = AutoTokenizer.from_pretrained(model_id)
generated_text = model.text_generation(
tokenizer=tokenizer,
prompt="Hello I am doing a project for my school",
max_seq_len=len(tokenizer.encode(EXPECTED_GENERATED_TEXT)),
)
self.assertEqual(generated_text, EXPECTED_GENERATED_TEXT)