Native Qwen2.5-VL vision encoder distributed with Lance.

Converts the officially separate Lance vision file while preserving tensor names, dtypes and bytes. Original window/full-attention schedule and 2x2 merging are retained. Image and multi-temporal-grid component tests are included. Lance shared MoT backbone and Wan2.2 VAE are not included; this is not the complete Lance generator.

Source and conversion

Model/weights: bytedance-research/Lance, fixed revision 7395315758865e6f56ab87ad06a88c7ac172f056. Native code: bytedance/Lance, fixed revision 4baeee086648996f6ab12e673cbe461b0b149997.

All selected tensor names, shapes, dtypes and payload bytes are preserved, except the explicit IBQ EMA shadow-name mapping. See extraction.json, provenance.json and tensor/conversion mappings. This repository contains component weights and a local native loader, not an AutoModel.from_pretrained checkpoint.

Load

Install requirements.txt in a compatible environment, then:

from huggingface_hub import snapshot_download
import sys
import torch
root = snapshot_download("gatilin/lance_qwen2_5_vl_vision")
sys.path.insert(0, root)
from component import load_model
model = load_model(root, dtype=torch.float32, device="cpu")

The model is returned in evaluation mode. Run inference under torch.inference_mode(). Inputs are native tensors/patches, not arbitrary PIL images; inspect component.py and the included validation script for the exact layout. CPU float32 was tested; CUDA/MPS and full end-to-end benchmarks were not tested.

Validation

See validation.json for the executed component tests, shapes and limitations. Source payload equality and native strict loading are required before release. Component tests do not establish full-model generation or understanding quality. Original stochastic decoding settings and normalization conventions must be retained.

License and attribution

Original upstream notices and license files are retained. Reused third-party components keep their original terms. This extraction does not claim authorship of the underlying trained model.