A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI][Bugfix] Skip Ovis2 generation test because of broken remote code (#22954)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Isotr0py committed
Aug 16, 2025 at 17:44 UTC
2dbccce8a67e8004b365e7e533107c54c9542ce7
1 file changed
+11
-4
tests/models/registry.py
+11
-4
@@ -196,7 +196,8 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
196
{"alias": "gpt2"}),
197
"GPTBigCodeForCausalLM": _HfExamplesInfo("bigcode/starcoder",
198
extras={"tiny": "bigcode/tiny_starcoder_py"}, # noqa: E501
199
- min_transformers_version="4.55.1"),
199
+ min_transformers_version="4.55.1",
200
+ transformers_version_reason="HF model broken in 4.55.0"), # noqa: E501
201
"GPTJForCausalLM": _HfExamplesInfo("Milos/slovak-gpt-j-405M",
202
{"6b": "EleutherAI/gpt-j-6b"}),
203
"GPTNeoXForCausalLM": _HfExamplesInfo("EleutherAI/pythia-70m",
@@ -408,14 +409,16 @@ _MULTIMODAL_EXAMPLE_MODELS = {
409
extras={"2b": "h2oai/h2ovl-mississippi-2b"}, # noqa: E501
410
max_transformers_version="4.48", # noqa: E501
411
transformers_version_reason="HF model is not compatible."), # noqa: E501
412
+ "Idefics3ForConditionalGeneration": _HfExamplesInfo("HuggingFaceM4/Idefics3-8B-Llama3", # noqa: E501
413
+ {"tiny": "HuggingFaceTB/SmolVLM-256M-Instruct"}, # noqa: E501
414
+ min_transformers_version="4.55.1",
415
+ transformers_version_reason="HF model broken in 4.55.0"), # noqa: E501
416
"InternVLChatModel": _HfExamplesInfo("OpenGVLab/InternVL2-1B",
417
extras={"2B": "OpenGVLab/InternVL2-2B",
418
"3.0": "OpenGVLab/InternVL3-1B"}, # noqa: E501
419
trust_remote_code=True),
420
"InternS1ForConditionalGeneration": _HfExamplesInfo("internlm/Intern-S1",
421
trust_remote_code=True),
417
- "Idefics3ForConditionalGeneration": _HfExamplesInfo("HuggingFaceM4/Idefics3-8B-Llama3", # noqa: E501
418
- {"tiny": "HuggingFaceTB/SmolVLM-256M-Instruct"}), # noqa: E501
422
"KeyeForConditionalGeneration": _HfExamplesInfo("Kwai-Keye/Keye-VL-8B-Preview", # noqa: E501
423
trust_remote_code=True),
424
"KimiVLForConditionalGeneration": _HfExamplesInfo("moonshotai/Kimi-VL-A3B-Instruct", # noqa: E501
@@ -455,6 +458,8 @@ _MULTIMODAL_EXAMPLE_MODELS = {
458
"Llama_Nemotron_Nano_VL" : _HfExamplesInfo("nvidia/Llama-3.1-Nemotron-Nano-VL-8B-V1", # noqa: E501
459
trust_remote_code=True),
460
"Ovis": _HfExamplesInfo("AIDC-AI/Ovis2-1B", trust_remote_code=True,
461
+ max_transformers_version="4.53",
462
+ transformers_version_reason="HF model is not compatible", # noqa: E501
463
extras={"1.6-llama": "AIDC-AI/Ovis1.6-Llama3.2-3B",
464
"1.6-gemma": "AIDC-AI/Ovis1.6-Gemma2-9B"}), # noqa: E501
465
"PaliGemmaForConditionalGeneration": _HfExamplesInfo("google/paligemma-3b-mix-224", # noqa: E501
@@ -482,7 +487,9 @@ _MULTIMODAL_EXAMPLE_MODELS = {
487
"Qwen2_5OmniForConditionalGeneration": _HfExamplesInfo("Qwen/Qwen2.5-Omni-7B-AWQ"), # noqa: E501
488
"SkyworkR1VChatModel": _HfExamplesInfo("Skywork/Skywork-R1V-38B",
489
trust_remote_code=True),
485
- "SmolVLMForConditionalGeneration": _HfExamplesInfo("HuggingFaceTB/SmolVLM2-2.2B-Instruct"), # noqa: E501
490
+ "SmolVLMForConditionalGeneration": _HfExamplesInfo("HuggingFaceTB/SmolVLM2-2.2B-Instruct", # noqa: E501
491
+ min_transformers_version="4.55.1",
492
+ transformers_version_reason="HF model broken in 4.55.0"), # noqa: E501
493
"Step3VLForConditionalGeneration": _HfExamplesInfo("stepfun-ai/step3",
494
trust_remote_code=True,
495
is_available_online=False),