xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-13 02:25:36 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI/Build] Skip Mllama HF runner tests with Transformers v4.55.0 (#22659)

Browse Source 
Signed-off-by: Isotr0py <2037008807@qq.com>
This commit is contained in:
Isotr0py 2025-08-12 01:00:58 +08:00 committed by GitHub
parent 84cf78acee
commit c90fb03df5
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
tests/models/multimodal/generation/test_mllama.py
View File

@ -6,6 +6,7 @@ from typing import Optional, overload
import pytest import pytest
import torch import torch
from transformers import AutoConfig, AutoModelForImageTextToText, AutoTokenizer from transformers import AutoConfig, AutoModelForImageTextToText, AutoTokenizer
from transformers import __version__ as TRANSFORMERS_VERSION
from vllm import LLM, SamplingParams from vllm import LLM, SamplingParams
from vllm.attention.backends.flash_attn import FlashAttentionMetadata from vllm.attention.backends.flash_attn import FlashAttentionMetadata
@ -285,6 +286,10 @@ def clear_cache():
@pytest.mark.parametrize("max_tokens", [128]) @pytest.mark.parametrize("max_tokens", [128])
@pytest.mark.parametrize("num_logprobs", [5]) @pytest.mark.parametrize("num_logprobs", [5])
@pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS) @pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS)
@pytest.mark.skipif(
TRANSFORMERS_VERSION == "4.55.0",
reason="Transformers v4.55.0 has a regression issue on mllama, "
"see: https://github.com/huggingface/transformers/pull/40083")
def test_models_single_leading_image(hf_runner, vllm_runner, image_assets, def test_models_single_leading_image(hf_runner, vllm_runner, image_assets,
model, sizes, dtype, max_tokens, model, sizes, dtype, max_tokens,
num_logprobs, num_logprobs,
@ -313,6 +318,10 @@ def test_models_single_leading_image(hf_runner, vllm_runner, image_assets,
@pytest.mark.parametrize("max_tokens", [128]) @pytest.mark.parametrize("max_tokens", [128])
@pytest.mark.parametrize("num_logprobs", [5]) @pytest.mark.parametrize("num_logprobs", [5])
@pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS) @pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS)
@pytest.mark.skipif(
TRANSFORMERS_VERSION == "4.55.0",
reason="Transformers v4.55.0 has a regression issue on mllama, "
"see: https://github.com/huggingface/transformers/pull/40083")
def test_models_multi_leading_images(hf_runner, vllm_runner, image_assets, def test_models_multi_leading_images(hf_runner, vllm_runner, image_assets,
model, dtype, max_tokens, num_logprobs, model, dtype, max_tokens, num_logprobs,
attn_backend: _Backend) -> None: attn_backend: _Backend) -> None:
@ -362,6 +371,10 @@ def test_models_multi_leading_images(hf_runner, vllm_runner, image_assets,
@pytest.mark.parametrize("max_tokens", [128]) @pytest.mark.parametrize("max_tokens", [128])
@pytest.mark.parametrize("num_logprobs", [5]) @pytest.mark.parametrize("num_logprobs", [5])
@pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS) @pytest.mark.parametrize("attn_backend", LIST_ENC_DEC_SUPPORTED_BACKENDS)
@pytest.mark.skipif(
TRANSFORMERS_VERSION == "4.55.0",
reason="Transformers v4.55.0 has a regression issue on mllama, "
"see: https://github.com/huggingface/transformers/pull/40083")
def test_models_interleaved_images(hf_runner, vllm_runner, image_assets, model, def test_models_interleaved_images(hf_runner, vllm_runner, image_assets, model,
dtype, max_tokens, num_logprobs, dtype, max_tokens, num_logprobs,
attn_backend: _Backend) -> None: attn_backend: _Backend) -> None:
@ -402,6 +415,10 @@ def test_models_interleaved_images(hf_runner, vllm_runner, image_assets, model,
@pytest.mark.parametrize("dtype", ["bfloat16"]) @pytest.mark.parametrize("dtype", ["bfloat16"])
@pytest.mark.parametrize("max_tokens", [64]) @pytest.mark.parametrize("max_tokens", [64])
@pytest.mark.parametrize("num_logprobs", [5]) @pytest.mark.parametrize("num_logprobs", [5])
@pytest.mark.skipif(
TRANSFORMERS_VERSION == "4.55.0",
reason="Transformers v4.55.0 has a regression issue on mllama, "
"see: https://github.com/huggingface/transformers/pull/40083")
def test_models_distributed( def test_models_distributed(
hf_runner, hf_runner,
vllm_runner, vllm_runner,