Using `Qwen/Qwen2-VL-2B-Instruct`as base

#3
by Cossale - opened

I'm really happy with how well this new version is working, but it's a bit slow. What if we try out the new Qwen 2 VL model that just came out yesterday? It's much smaller than phi 3.5 vision, so it might be faster. Also hoping for multi-lingual support. Thanks for awesome release!

Owner

I'm really happy with how well this new version is working, but it's a bit slow. What if we try out the new Qwen 2 VL model that just came out yesterday? It's much smaller than phi 3.5 vision, so it might be faster. Also hoping for multi-lingual support. Thanks for awesome release!

@Cossale Thank you! Smaller versions are being trained and tested right now.

Sign up or log in to comment