huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-11 12:24:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

add integration tests

Browse Source
This commit is contained in:
OlivierDehaene 2024-02-20 16:09:21 +01:00
parent 5da4b01cc5
commit 767d21cbf8
6 changed files with 599 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
integration-tests/conftest.py
View File

@ -40,6 +40,9 @@ class ResponseComparator(JSONSnapshotExtension):
exclude=None,
matcher=None,
):
if isinstance(data, Response):
data = data.dict()
if isinstance(data, List):
data = [d.dict() for d in data]

89
integration-tests/models/__snapshots__/test_flash_gemma/test_flash_gemma.json Normal file
View File

@ -0,0 +1,89 @@
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": null,
"tokens": [
{
"id": 1736,
"logprob": -2.09375,
"special": false,
"text": " form"
},
{
"id": 109,
"logprob": -1.8671875,
"special": false,
"text": "\n\n"
},
{
"id": 651,
"logprob": -2.4375,
"special": false,
"text": "The"
},
{
"id": 2121,
"logprob": -1.8203125,
"special": false,
"text": " test"
},
{
"id": 3853,
"logprob": -0.23242188,
"special": false,
"text": " request"
},
{
"id": 1736,
"logprob": -0.08544922,
"special": false,
"text": " form"
},
{
"id": 603,
"logprob": -0.9375,
"special": false,
"text": " is"
},
{
"id": 1671,
"logprob": -1.671875,
"special": false,
"text": " used"
},
{
"id": 577,
"logprob": -0.40429688,
"special": false,
"text": " to"
},
{
"id": 3853,
"logprob": -1.1875,
"special": false,
"text": " request"
}
],
"top_tokens": null
},
"generated_text": " form\n\nThe test request form is used to request"
}

89
integration-tests/models/__snapshots__/test_flash_gemma/test_flash_gemma_all_params.json Normal file
View File

@ -0,0 +1,89 @@
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": 0,
"tokens": [
{
"id": 7539,
"logprob": -0.73046875,
"special": false,
"text": " forms"
},
{
"id": 708,
"logprob": 0.0,
"special": false,
"text": " are"
},
{
"id": 671,
"logprob": -1.703125,
"special": false,
"text": " an"
},
{
"id": 8727,
"logprob": 0.0,
"special": false,
"text": " essential"
},
{
"id": 1702,
"logprob": 0.0,
"special": false,
"text": " part"
},
{
"id": 576,
"logprob": 0.0,
"special": false,
"text": " of"
},
{
"id": 573,
"logprob": 0.0,
"special": false,
"text": " the"
},
{
"id": 11859,
"logprob": -1.6953125,
"special": false,
"text": " lab"
},
{
"id": 2185,
"logprob": -1.3125,
"special": false,
"text": " process"
},
{
"id": 578,
"logprob": -1.5,
"special": false,
"text": " and"
}
],
"top_tokens": null
},
"generated_text": "Test request forms are an essential part of the lab process and"
}

358
integration-tests/models/__snapshots__/test_flash_gemma/test_flash_gemma_load.json Normal file
View File

@ -0,0 +1,358 @@
[
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": null,
"tokens": [
{
"id": 1736,
"logprob": -2.09375,
"special": false,
"text": " form"
},
{
"id": 109,
"logprob": -1.9140625,
"special": false,
"text": "\n\n"
},
{
"id": 651,
"logprob": -2.453125,
"special": false,
"text": "The"
},
{
"id": 2121,
"logprob": -1.8984375,
"special": false,
"text": " test"
},
{
"id": 3853,
"logprob": -0.23535156,
"special": false,
"text": " request"
},
{
"id": 1736,
"logprob": -0.091308594,
"special": false,
"text": " form"
},
{
"id": 603,
"logprob": -0.96875,
"special": false,
"text": " is"
},
{
"id": 1671,
"logprob": -1.6484375,
"special": false,
"text": " used"
},
{
"id": 577,
"logprob": -0.43164062,
"special": false,
"text": " to"
},
{
"id": 3853,
"logprob": -1.2421875,
"special": false,
"text": " request"
}
],
"top_tokens": null
},
"generated_text": " form\n\nThe test request form is used to request"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": null,
"tokens": [
{
"id": 1736,
"logprob": -2.09375,
"special": false,
"text": " form"
},
{
"id": 109,
"logprob": -1.9140625,
"special": false,
"text": "\n\n"
},
{
"id": 651,
"logprob": -2.453125,
"special": false,
"text": "The"
},
{
"id": 2121,
"logprob": -1.8984375,
"special": false,
"text": " test"
},
{
"id": 3853,
"logprob": -0.23535156,
"special": false,
"text": " request"
},
{
"id": 1736,
"logprob": -0.091308594,
"special": false,
"text": " form"
},
{
"id": 603,
"logprob": -0.96875,
"special": false,
"text": " is"
},
{
"id": 1671,
"logprob": -1.6484375,
"special": false,
"text": " used"
},
{
"id": 577,
"logprob": -0.43164062,
"special": false,
"text": " to"
},
{
"id": 3853,
"logprob": -1.2421875,
"special": false,
"text": " request"
}
],
"top_tokens": null
},
"generated_text": " form\n\nThe test request form is used to request"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": null,
"tokens": [
{
"id": 1736,
"logprob": -2.09375,
"special": false,
"text": " form"
},
{
"id": 109,
"logprob": -1.9140625,
"special": false,
"text": "\n\n"
},
{
"id": 651,
"logprob": -2.453125,
"special": false,
"text": "The"
},
{
"id": 2121,
"logprob": -1.8984375,
"special": false,
"text": " test"
},
{
"id": 3853,
"logprob": -0.23535156,
"special": false,
"text": " request"
},
{
"id": 1736,
"logprob": -0.091308594,
"special": false,
"text": " form"
},
{
"id": 603,
"logprob": -0.96875,
"special": false,
"text": " is"
},
{
"id": 1671,
"logprob": -1.6484375,
"special": false,
"text": " used"
},
{
"id": 577,
"logprob": -0.43164062,
"special": false,
"text": " to"
},
{
"id": 3853,
"logprob": -1.2421875,
"special": false,
"text": " request"
}
],
"top_tokens": null
},
"generated_text": " form\n\nThe test request form is used to request"
},
{
"details": {
"best_of_sequences": null,
"finish_reason": "length",
"generated_tokens": 10,
"prefill": [
{
"id": 2,
"logprob": null,
"text": "<bos>"
},
{
"id": 2015,
"logprob": -10.0,
"text": "Test"
},
{
"id": 3853,
"logprob": -10.875,
"text": " request"
}
],
"seed": null,
"tokens": [
{
"id": 1736,
"logprob": -2.09375,
"special": false,
"text": " form"
},
{
"id": 109,
"logprob": -1.9140625,
"special": false,
"text": "\n\n"
},
{
"id": 651,
"logprob": -2.453125,
"special": false,
"text": "The"
},
{
"id": 2121,
"logprob": -1.8984375,
"special": false,
"text": " test"
},
{
"id": 3853,
"logprob": -0.23535156,
"special": false,
"text": " request"
},
{
"id": 1736,
"logprob": -0.091308594,
"special": false,
"text": " form"
},
{
"id": 603,
"logprob": -0.96875,
"special": false,
"text": " is"
},
{
"id": 1671,
"logprob": -1.6484375,
"special": false,
"text": " used"
},
{
"id": 577,
"logprob": -0.43164062,
"special": false,
"text": " to"
},
{
"id": 3853,
"logprob": -1.2421875,
"special": false,
"text": " request"
}
],
"top_tokens": null
},
"generated_text": " form\n\nThe test request form is used to request"
}
]

58
integration-tests/models/test_flash_gemma.py Normal file
View File

@ -0,0 +1,58 @@
import pytest
@pytest.fixture(scope="module")
def flash_gemma_handle(launcher):
with launcher("gg-hf/gemma-2b", num_shard=1) as handle:
yield handle
@pytest.fixture(scope="module")
async def flash_gemma(flash_gemma_handle):
await flash_gemma_handle.health(300)
return flash_gemma_handle.client
@pytest.mark.asyncio
@pytest.mark.private
async def test_flash_gemma(flash_gemma, response_snapshot):
response = await flash_gemma.generate(
"Test request", max_new_tokens=10, decoder_input_details=True
)
assert response.details.generated_tokens == 10
assert response == response_snapshot
@pytest.mark.asyncio
@pytest.mark.private
async def test_flash_gemma_all_params(flash_gemma, response_snapshot):
response = await flash_gemma.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_gemma_load(flash_gemma, generate_load, response_snapshot):
responses = await generate_load(flash_gemma, "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

4
server/text_generation_server/models/custom_modeling/flash_gemma_modeling.py
View File

@ -68,8 +68,6 @@ answers should not include any harmful, unethical, racist, sexist, toxic, danger
If a question does not make any sense, or is not factually coherent, explain why instead of answering something not \
correct. If you don't know the answer to a question, please don't share false information."""
# fmt: on
@ -185,12 +183,10 @@ class GemmaTokenizerFast(PreTrainedTokenizerFast):
return (out_vocab_file,)
@property
# Copied from transformers.models.llama.tokenization_llama.GemmaTokenizer.default_chat_template
def default_chat_template(self):
raise NotImplementedError
# TODO ArthurZ let's rely on the template processor instead, refactor all fast tokenizers
# Copied from transformers.models.llama.tokenization_llama.GemmaTokenizer.build_inputs_with_special_tokens
def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
bos_token_id = [self.bos_token_id] if self.add_bos_token else []
eos_token_id = [self.eos_token_id] if self.add_eos_token else []