Downloads · 30 days
9
17% of all-time downloads
Irisssme/EAGLE-2-llava-v1.6-vicuna-7b-hf
EAGLE-2-llava-v1.6-vicuna-7b-hf is a machine learning model from Irisssme. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
This repository contains an EAGLE-2 drafter checkpoint for llava-hf/llava-v1.6-vicuna-7b-hf.
Downloads · 30 days
9
17% of all-time downloads
All-time downloads
53
Public
Parameters
367M
1.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors1.5 GB · 100%
From the Hugging Face model README
This repository contains an EAGLE-2 drafter checkpoint for llava-hf/llava-v1.6-vicuna-7b-hf.
This checkpoint is not a standalone vision-language model. It must be used with the corresponding base model and speculative-decoding inference code. The checkpoint consumes the complete multimodal sequence and does not prune visual tokens.
Install the HiViS codebase, then run:
python -m hivis.evaluation.ge_hivis_answer \
--draft-method eagle \
--base-model-path llava-hf/llava-v1.6-vicuna-7b-hf \
--ea-model-path Irisssme/EAGLE-2-llava-v1.6-vicuna-7b-hf \
--dataset <dataset> \
--answer-file outputs/eagle.jsonl
config.json: drafter configuration.model.safetensors: EAGLE-2 drafter weights.This checkpoint and its inference path build on the foundational work of EAGLE. Please cite the original EAGLE and EAGLE-2 papers when using this checkpoint.
This checkpoint is released under the Apache License 2.0. Users must also follow the license of the corresponding base model and the upstream EAGLE project.