Zi-Yuan Hu

Why not compare with FLAN-T5-Large?

2

What considerations preclude a comparison against the FLAN-T5-Large model? As shown in MM-COT paper (arXiv:2302.00923) Table7, FLAN-T5-Large achieves ~93% on ScienceQA. Is it because the experimental settings are different? By...

area: 视觉和语言(Vision-Language) title: VL-PET: Vision-and-Language Parameter-Efficient Tuning via Granularity Control paper: https://arxiv.org/abs/2308.09804 code: https://github.com/HenryHZY/VL-PET project page: https://henryhzy.github.io/VL-PET

Instruction about training Open-Flamingo from scratch

26

Hi @anas-awadalla As described in #124, "Our training took place on 32 80GB A100s. We trained on 5M samples from MMC4 and 10M from LAION 2B." I am interested in...

How to use Qwen3-VL generate() with num_return_sequences > 1?

1

Dear Qwen Team, I would like to ask how can we use Qwen3-VL generate() with num_return_sequences > 1? Here is my code with num_return_sequences > 1? ``` from transformers import...