-
The GPU memory consumption of the model was too high, so I converted it to a LLAMA.CPP file. The GPU memory usage is fine.
However, due to the nature of the model converted to llama.cpp in the model …
-
Hi, thanks for your amazing work! how do you split train, val and test set of ocr-vqa dataset, there is not split method on official webset.
-
Hi! I have some questions about the files preprocessed from the dataset VQA-CP v2 and used in your work. Because VQA-CP v2 only has the training and test questions and annotations, how to generate fil…
-
### Describe the issue
Issue: I find download ocr vqa images are slow and I implement a multi-thread download file.
Command:
```
import concurrent.futures
def download_image(k):
ext = os.p…
-
Hello,
I've been deploying my VQA (Vision Query Answer) model using Triton Server and utilizing the `perf_analyzer` tool for benchmarking. However, using random data for the VQA model leads to unde…
-
Dear authors,
Thanks for your excellent work! But when I follow the VQA inference steps, I encounter the following problems.
1. When I use scripts `evaluate_vqa_allcand.sh` and `evaluate_vqa_allca…
-
### Question
I looked at the eval script
```
CONV="conv_template"
CKPT_NAME="your_ckpt_name"
CKPT="checkpoints/${CKPT_NAME}"
EVAL="eval"
deepspeed moellava/eval/model_vqa_loader.py \
--m…
-
We've run into an extremely weird series of bugs with Charts.js in the VQA feature. Firstly, the chart "randomly" reordered the association of bar chart labels to bar chart data values. Through experi…
-
Hi, when will the weights for blip-2 fine-tuned for VQA be released? We've been waiting for quite some time. I'd really appreciate it if they can be released as soon as possible. :) Thanks!
-
I could not access the dataset. could you please provide more instructions since it is not uploaded in English portal?