[Multimodal] Consolidate mm inputs into MultiModalFeatureSpec (#23779)

Signed-off-by: sfeng33 <4florafeng@gmail.com>
This commit is contained in:
Flora Feng
2025-08-29 18:36:57 +08:00
committed by GitHub
parent d9e00dbd1f
commit 69f46359dd
16 changed files with 143 additions and 146 deletions
+1 -3
View File
@@ -35,9 +35,7 @@ def make_request() -> EngineCoreRequest:
return EngineCoreRequest(
request_id=str(uuid.uuid4()),
prompt_token_ids=PROMPT_TOKENS,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
sampling_params=SamplingParams(),
pooling_params=None,
eos_token_id=None,
+1 -3
View File
@@ -52,9 +52,7 @@ def make_request(
return EngineCoreRequest(
request_id=str(uuid.uuid4()),
prompt_token_ids=prompt_tokens_ids,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
sampling_params=params,
pooling_params=None,
eos_token_id=None,
+8 -10
View File
@@ -26,16 +26,14 @@ def test_fast_inc_detok_invalid_utf8_err_case():
prompt_token_ids = [107, 4606, 236787, 107]
params = SamplingParams(skip_special_tokens=True)
request = EngineCoreRequest(
"test",
prompt_token_ids,
None,
None,
None,
params,
None,
None,
0.0,
None,
request_id="test",
prompt_token_ids=prompt_token_ids,
mm_features=None,
sampling_params=params,
pooling_params=None,
eos_token_id=None,
arrival_time=0.0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,
)
+10 -20
View File
@@ -52,11 +52,9 @@ def test_incremental_detokenization(request_output_kind: RequestOutputKind,
requests = [
EngineCoreRequest(request_id=f"request-{idx}",
prompt_token_ids=prompt_tokens,
arrival_time=0,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
eos_token_id=None,
arrival_time=0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,
@@ -401,11 +399,9 @@ def test_logprobs_processor(request_output_kind: RequestOutputKind,
requests = [
EngineCoreRequest(request_id=request_id_list[idx],
prompt_token_ids=prompt_tokens,
arrival_time=0,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
eos_token_id=None,
arrival_time=0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,
@@ -566,11 +562,9 @@ def test_stop_token(include_stop_str_in_output: bool,
request = EngineCoreRequest(
request_id=request_id,
prompt_token_ids=prompt_tokens,
arrival_time=0,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
eos_token_id=eos_token_id,
arrival_time=0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,
@@ -665,11 +659,9 @@ def test_stop_string(include_stop_str_in_output: bool,
EngineCoreRequest(
request_id=request_id_list[idx],
prompt_token_ids=prompt_tokens,
arrival_time=0,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
eos_token_id=None,
arrival_time=0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,
@@ -781,11 +773,9 @@ def test_iteration_stats(dummy_test_vectors):
EngineCoreRequest(
request_id=f"request-{idx}",
prompt_token_ids=prompt_tokens,
arrival_time=0,
mm_kwargs=None,
mm_hashes=None,
mm_placeholders=None,
mm_features=None,
eos_token_id=None,
arrival_time=0,
lora_request=None,
cache_salt=None,
data_parallel_rank=None,