xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-07-09 20:07:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

Support for NemotronH Nano VLM (#23644)

Browse Source 
Signed-off-by: Daniel Afrimi <danielafrimi8@gmail.com>
This commit is contained in:
danielafrimi 2025-09-10 16:10:06 +03:00 committed by GitHub
parent 8b83b93739
commit 72d30108a0
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 1400 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
tests/models/registry.py
View File

@ -515,6 +515,9 @@ _MULTIMODAL_EXAMPLE_MODELS = {
trust_remote_code=True), trust_remote_code=True),
"Llama_Nemotron_Nano_VL" : _HfExamplesInfo("nvidia/Llama-3.1-Nemotron-Nano-VL-8B-V1", # noqa: E501 "Llama_Nemotron_Nano_VL" : _HfExamplesInfo("nvidia/Llama-3.1-Nemotron-Nano-VL-8B-V1", # noqa: E501
trust_remote_code=True), trust_remote_code=True),
"NemotronH_Nano_VL": _HfExamplesInfo("nano_vl_dummy",
is_available_online=False,
trust_remote_code=True),
"Ovis": _HfExamplesInfo("AIDC-AI/Ovis2-1B", trust_remote_code=True, "Ovis": _HfExamplesInfo("AIDC-AI/Ovis2-1B", trust_remote_code=True,
max_transformers_version="4.53", max_transformers_version="4.53",
transformers_version_reason="HF model is not compatible", # noqa: E501 transformers_version_reason="HF model is not compatible", # noqa: E501

2
vllm/config/__init__.py
View File

@ -1552,7 +1552,7 @@ class ModelConfig:
for bc in block_configs[start:end]) for bc in block_configs[start:end])
else: else:
# Hybrid model Jamba # Hybrid model Jamba
layers_block_type_value = getattr(self.hf_config, layers_block_type_value = getattr(self.hf_text_config,
"layers_block_type", None) "layers_block_type", None)
if layers_block_type_value is not None: if layers_block_type_value is not None:
if hasattr(self.hf_text_config, if hasattr(self.hf_text_config,

1395
vllm/model_executor/models/nano_nemotron_vl.py Normal file
View File

File diff suppressed because it is too large Load Diff

1
vllm/model_executor/models/registry.py
View File

@ -223,6 +223,7 @@ _MULTIMODAL_MODELS = {
"GraniteSpeechForConditionalGeneration": ("granite_speech", "GraniteSpeechForConditionalGeneration"), # noqa: E501 "GraniteSpeechForConditionalGeneration": ("granite_speech", "GraniteSpeechForConditionalGeneration"), # noqa: E501
"H2OVLChatModel": ("h2ovl", "H2OVLChatModel"), "H2OVLChatModel": ("h2ovl", "H2OVLChatModel"),
"InternVLChatModel": ("internvl", "InternVLChatModel"), "InternVLChatModel": ("internvl", "InternVLChatModel"),
"NemotronH_Nano_VL": ("nano_nemotron_vl", "NemotronH_Nano_VL"),
"InternS1ForConditionalGeneration": ("interns1", "InternS1ForConditionalGeneration"), # noqa: E501 "InternS1ForConditionalGeneration": ("interns1", "InternS1ForConditionalGeneration"), # noqa: E501
"InternVLForConditionalGeneration": ("interns1", "InternS1ForConditionalGeneration"), # noqa: E501 "InternVLForConditionalGeneration": ("interns1", "InternS1ForConditionalGeneration"), # noqa: E501
"Idefics3ForConditionalGeneration":("idefics3","Idefics3ForConditionalGeneration"), "Idefics3ForConditionalGeneration":("idefics3","Idefics3ForConditionalGeneration"),