bridge.models.qwen_vl.qwen25_vl_bridge#

Module Contents#

Classes#

Qwen25VLBridge

Megatron Bridge for Qwen2.5-VL Conditional Generation.

API#

class bridge.models.qwen_vl.qwen25_vl_bridge.Qwen25VLBridge#

Bases: megatron.bridge.models.conversion.model_bridge.MegatronModelBridge

Megatron Bridge for Qwen2.5-VL Conditional Generation.

This bridge handles the conversion between HuggingFace Qwen2_5_VLForConditionalGeneration and Megatron-Core GPTModel formats, including weight mappings and configuration translation for vision-language models.

.. rubric:: Example

from megatron.bridge import AutoBridge bridge = AutoBridge.from_hf_pretrained(“Qwen/Qwen2.5-VL-3B-Instruct”) provider = bridge.to_megatron_provider()

provider_bridge(
hf_pretrained: megatron.bridge.models.hf_pretrained.vlm.PreTrainedVLM,
) → megatron.bridge.models.qwen_vl.qwen_vl_provider.Qwen25VLModelProvider#
mapping_registry() → megatron.bridge.models.conversion.mapping_registry.MegatronMappingRegistry#