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 11, 13:45 UTC
[CI] Update several models in registry that are available online now (#30514)
Signed-off-by: mgoin <mgoin64@gmail.com> Signed-off-by: Michael Goin <mgoin64@gmail.com> Co-authored-by: Isotr0py <2037008807@qq.com>
Michael Goin committed
Dec 12, 2025 at 21:28 UTC
2f32a68d75324299d13025c75f9cb5427e5c445d
2 files changed
+6
-6
.buildkite/test-pipeline.yaml
+2
@@ -692,6 +692,7 @@ steps:
692
source_file_dependencies:
693
- vllm/
694
- tests/models/test_initialization.py
695
+ - tests/models/registry.py
696
commands:
697
# Run a subset of model initialization tests
698
- pytest -v -s models/test_initialization.py::test_can_initialize_small_subset
@@ -704,6 +705,7 @@ steps:
705
- vllm/model_executor/models/
706
- vllm/transformers_utils/
707
- tests/models/test_initialization.py
708
+ - tests/models/registry.py
709
commands:
710
# Only when vLLM model source is modified - test initialization of a large
711
# subset of supported models (the complement of the small subset in the above
tests/models/registry.py
+4
-6
@@ -356,7 +356,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
356
),
357
"MistralForCausalLM": _HfExamplesInfo("mistralai/Mistral-7B-Instruct-v0.1"),
358
"MistralLarge3ForCausalLM": _HfExamplesInfo(
359
- "mistralai/Mistral-Large-3-675B-Instruct-2512-NVFP4", is_available_online=False
359
+ "mistralai/Mistral-Large-3-675B-Instruct-2512-NVFP4"
360
),
361
"MixtralForCausalLM": _HfExamplesInfo(
362
"mistralai/Mixtral-8x7B-Instruct-v0.1",
@@ -635,7 +635,7 @@ _MULTIMODAL_EXAMPLE_MODELS = {
635
),
636
"HunYuanVLForConditionalGeneration": _HfExamplesInfo(
637
"tencent/HunyuanOCR",
638
- is_available_online=False,
638
+ hf_overrides={"num_experts": 0},
639
),
640
"Idefics3ForConditionalGeneration": _HfExamplesInfo(
641
"HuggingFaceM4/Idefics3-8B-Llama3",
@@ -674,8 +674,7 @@ _MULTIMODAL_EXAMPLE_MODELS = {
674
"https://huggingface.co/moonshotai/Kimi-VL-A3B-Instruct/discussions/31",
675
),
676
"LightOnOCRForConditionalGeneration": _HfExamplesInfo(
677
- "lightonai/LightOnOCR-1B",
678
- is_available_online=False,
677
+ "lightonai/LightOnOCR-1B-1025"
678
),
679
"Llama4ForConditionalGeneration": _HfExamplesInfo(
680
"meta-llama/Llama-4-Scout-17B-16E-Instruct",
@@ -779,8 +778,6 @@ _MULTIMODAL_EXAMPLE_MODELS = {
778
"ministral-3": "mistralai/Ministral-3-3B-Instruct-2512",
779
},
780
tokenizer_mode="mistral",
782
- # TODO: revert once Mistral-Large-3 and Ministral-3 are publicly available.
783
- is_available_online=False,
781
),
782
"QwenVLForConditionalGeneration": _HfExamplesInfo(
783
"Qwen/Qwen-VL",
@@ -886,6 +883,7 @@ _SPECULATIVE_DECODING_EXAMPLE_MODELS = {
883
"EagleMistralLarge3ForCausalLM": _HfExamplesInfo(
884
"mistralai/Mistral-Large-3-675B-Instruct-2512",
885
speculative_model="mistralai/Mistral-Large-3-675B-Instruct-2512-Eagle",
886
+ # TODO: revert once figuring out OOM in CI
887
is_available_online=False,
888
),
889
"LlamaForCausalLMEagle3": _HfExamplesInfo(