EAGLE-Qwen2.5-VL-7B-Instruct

This repository provides the checkpoint used by both EAGLE1 and EAGLE2 in our MMSpec experiments.

Important Note

  • EAGLE1 and EAGLE2 share the same checkpoint files in this repo.
  • The two files below are weight shards of one checkpoint, not two different models:
    • model.safetensors
    • model_1.safetensors

Why two .safetensors files?

The checkpoint was saved in sharded format due to model/file size. Most loading code will automatically handle both shards together as a single model checkpoint.

Scope

This model card mainly clarifies checkpoint sharing between EAGLE1 and EAGLE2.

If you use this checkpoint in your project, please cite the corresponding EAGLE/MMSpec work.

Downloads last month
32
Safetensors
Model size
0.8B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Collection including Cloudriver/EAGLE-Qwen2.5-VL-7B-Instruct