mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-05-20 17:22:09 +00:00
7e531f413d
* Update to Torch 2.7.0 * Try to fix typer/click issue * Pin click to fix incompatibility with typer * Fix some test outputs with slight deviations * Attempt again to sync with CI * Mamba too * Fixup mllama Also switch to `unsloth/Llama-3.2-11B-Vision-Instruct` for testing from the EU :).
29 lines
818 B
Python
29 lines
818 B
Python
import pytest
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def flash_pali_gemma_handle(launcher):
|
|
with launcher(
|
|
"google/paligemma2-3b-pt-224",
|
|
) as handle:
|
|
yield handle
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
async def flash_pali_gemma(flash_pali_gemma_handle):
|
|
await flash_pali_gemma_handle.health(300)
|
|
return flash_pali_gemma_handle.client
|
|
|
|
|
|
async def test_flash_pali_gemma_image(flash_pali_gemma, response_snapshot):
|
|
car_image = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/car.jpg"
|
|
response = await flash_pali_gemma.generate(
|
|
f"",
|
|
max_new_tokens=20,
|
|
)
|
|
assert (
|
|
response.generated_text == "\nBrown\nCar\nColor\nCool\nDecor\n\n\n\n\n\n\n?\n?"
|
|
)
|
|
|
|
assert response == response_snapshot
|