mirror of
https://github.com/huggingface/text-generation-inference.git
synced 2025-04-19 13:52:07 +00:00
67ef0649cf
* Add more representative Llama GPTQ test The Llama GPTQ test is updated to use a model with the commonly-used quantizer config format and activation sorting. The old test is kept around (but renamed) since it tests the format produced by `text-generation-server quantize`. * Add support for manually triggering a release build
67 lines
1.7 KiB
Python
67 lines
1.7 KiB
Python
import pytest
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def flash_llama_gptq_handle(launcher):
|
|
with launcher(
|
|
"astronomer/Llama-3-8B-Instruct-GPTQ-4-Bit", num_shard=2, quantize="gptq"
|
|
) as handle:
|
|
yield handle
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
async def flash_llama_gptq(flash_llama_gptq_handle):
|
|
await flash_llama_gptq_handle.health(300)
|
|
return flash_llama_gptq_handle.client
|
|
|
|
|
|
@pytest.mark.release
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_llama_gptq(flash_llama_gptq, response_snapshot):
|
|
response = await flash_llama_gptq.generate(
|
|
"Test request", max_new_tokens=10, decoder_input_details=True
|
|
)
|
|
|
|
assert response.details.generated_tokens == 10
|
|
assert response == response_snapshot
|
|
|
|
|
|
@pytest.mark.release
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_llama_gptq_all_params(flash_llama_gptq, response_snapshot):
|
|
response = await flash_llama_gptq.generate(
|
|
"Test request",
|
|
max_new_tokens=10,
|
|
repetition_penalty=1.2,
|
|
return_full_text=True,
|
|
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.release
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.private
|
|
async def test_flash_llama_gptq_load(
|
|
flash_llama_gptq, generate_load, response_snapshot
|
|
):
|
|
responses = await generate_load(
|
|
flash_llama_gptq, "Test request", max_new_tokens=10, n=4
|
|
)
|
|
|
|
assert len(responses) == 4
|
|
assert all([r.generated_text == responses[0].generated_text for r in responses])
|
|
|
|
assert responses == response_snapshot
|