mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-04-19 22:02:06 +00:00
a6a0c97ed9
* wip * rollback * refactor to use prefix/postfix namming + fix all_input_ids_tensor * maybe patching vlms? * fix filter and concat * wip, no filter, no concat * current * add prepare_for_prefill * working * load tested * re-create slots * re-create slots * fix slot_filtering_indices * feedback loop * remove log * fix benchmarker * fix vlm and seq2seq * rename to cache and input lengths * fix prefill logprobs * fix launcher * fix logprobs? * idk at this point * max input length * omfg * remove debugging lines * fix tests * fix mllama * fix cargo tests * remove support chunking for paged * Fixing non blocked attentions * Fixing dtype + AMD, Ipex targets. * lint fix. * rename * Fix prefix_caching variable, remove defaults in server (confusing a lot of the times). * Add simple resolution when user specifies ATTENTION=paged. * Put back non default simple tests. * Fix env name --------- Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
91 lines
2.7 KiB
Python
91 lines
2.7 KiB
Python
import pytest
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def flash_idefics2_next_handle(launcher):
|
|
with launcher(
|
|
"HuggingFaceM4/idefics2-8b",
|
|
) as handle:
|
|
yield handle
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
async def flash_idefics2_next(flash_idefics2_next_handle):
|
|
await flash_idefics2_next_handle.health(300)
|
|
return flash_idefics2_next_handle.client
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_idefics2_next_simple(
|
|
flash_idefics2_next, response_snapshot, chicken
|
|
):
|
|
response = await flash_idefics2_next.generate(
|
|
f"User:Write me a short story<end_of_utterance> \nAssistant:",
|
|
max_new_tokens=10,
|
|
)
|
|
assert (
|
|
response.generated_text == " A chicken is sitting on a pile of money."
|
|
), f"{repr(response.generated_text)}"
|
|
assert response.details.generated_tokens == 10
|
|
assert response == response_snapshot
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_idefics2_two_images(
|
|
flash_idefics2_next, response_snapshot, chicken, cow_beach
|
|
):
|
|
response = await flash_idefics2_next.generate(
|
|
f"User:Where are the cow and chicken?<end_of_utterance> \nAssistant:",
|
|
max_new_tokens=20,
|
|
)
|
|
assert (
|
|
response.generated_text
|
|
== " The cow is standing on the beach and the chicken is sitting on a pile of money."
|
|
), f"{repr(response.generated_text)}"
|
|
assert response.details.generated_tokens == 19
|
|
assert response == response_snapshot
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_idefics2_next_all_params(flash_idefics2_next, response_snapshot):
|
|
response = await flash_idefics2_next.generate(
|
|
"Test request",
|
|
max_new_tokens=10,
|
|
repetition_penalty=1.2,
|
|
return_full_text=True,
|
|
stop_sequences=["test"],
|
|
temperature=0.5,
|
|
top_p=0.9,
|
|
top_k=10,
|
|
truncate=5,
|
|
typical_p=0.9,
|
|
watermark=True,
|
|
decoder_input_details=True,
|
|
seed=0,
|
|
)
|
|
|
|
assert response.details.generated_tokens == 10
|
|
assert response == response_snapshot
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_idefics2_next_load(
|
|
flash_idefics2_next, generate_load, response_snapshot, chicken
|
|
):
|
|
responses = await generate_load(
|
|
flash_idefics2_next,
|
|
f"User:Write me a short story<end_of_utterance> \nAssistant:",
|
|
max_new_tokens=10,
|
|
n=4,
|
|
)
|
|
generated_texts = [r.generated_text for r in responses]
|
|
assert generated_texts[0] == " A chicken is sitting on a pile of money."
|
|
assert len(generated_texts) == 4
|
|
assert all([r.generated_text == generated_texts[0] for r in responses])
|
|
|
|
assert responses == response_snapshot
|