apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 17:17:14 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

polish code with new pre-commit rule (#2923)

Browse Source
This commit is contained in:
Zero Rains
2025-07-19 23:19:27 +08:00
committed by GitHub
parent b8676d71a8
commit 25698d56d1
424 changed files with 14307 additions and 13518 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
test/operators/test_air_topp_sampling.py
View File

@@ -11,7 +11,7 @@
# 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.
""" UT for air_topp_sampling kernel """
"""UT for air_topp_sampling kernel"""
import subprocess
import unittest
@@ -23,7 +23,6 @@ import fastdeploy.model_executor.ops.gpu
class Test(unittest.TestCase):
def setUp(self):
"""
Initialize.
@@ -32,8 +31,7 @@ class Test(unittest.TestCase):
np.random.seed(42)
print(paddle.device.cuda.get_device_properties())
print(paddle.__git_commit__)
nvcc_output = subprocess.check_output(["nvcc", "--version"],
universal_newlines=True)
nvcc_output = subprocess.check_output(["nvcc", "--version"], universal_newlines=True)
output = nvcc_output.split()
release_idx = output.index("release") + 1
self.nvcc_cuda_version = float(output[release_idx].split(",")[0])
@@ -49,15 +47,15 @@ class Test(unittest.TestCase):
x = paddle.randn([bsz, vocab_size])
x = paddle.nn.functional.softmax(x)
x = paddle.cast(x, "float32")
top_ps = paddle.to_tensor(
np.random.uniform(0, 1, [bsz]).astype(np.float32))
top_ps = paddle.to_tensor(np.random.uniform(0, 1, [bsz]).astype(np.float32))
_, next_tokens = fastdeploy.model_executor.ops.gpu.air_topp_sampling(
x.cuda(), top_ps.cuda(), None, None, seed=0, k=1, mode="truncated")
x.cuda(), top_ps.cuda(), None, None, seed=0, k=1, mode="truncated"
)
print(next_tokens)
less_than_zero = next_tokens >= 0
greater_than_vocab_size = next_tokens <= vocab_size
accuracy = paddle.logical_and(less_than_zero, greater_than_vocab_size)
print(f'Accuracy of results: {accuracy}')
print(f"Accuracy of results: {accuracy}")
if __name__ == "__main__":