LlamaGen

LlamaGen copied to clipboard

May I ask why VQVAE's learned codebook was not used as the Embedding layer for autoregressive training of language models, but only the indices of the codebook were used, and...

wangyf8848

Hi, in the paper, you state "The codebook usage is calculated as the percentage of used codes in the queue of size 65536 over the whole codebook size." May I...

BaohaoLiao

Why is the model GPT in the code？

1

Hello, the paper mentions using Llama for autoregressive training, but why is the language model in the code using GPT ？

wangyf8848

When I open the `https://huggingface.co/spaces/FoundationVision/LlamaGen`, It is showed as follow: runtime error Traceback (most recent call last): File "/home/user/app/app.py", line 171, in t5_model, vq_model, gpt_model, image_size = load_model(args) File "/home/user/app/app.py",...

bigbrother001

T2I performance on mscoco

1

Hi, thanks for you job, could you show us how your model performs on MS-COCO?

HalvesChen

Hi~ Thank you for your work. I would like to know if you have compared the ablation experiments on the LAION-COCO fine-tuning.

HalvesChen

Hi, I am trying to finetune the decoder only of the tokenizer on a new dataset. I was wondering if you could share some finetuning recipies. Mainly: - Should the...

elias-ramzi

VQ-VAE ckpt optimizer states?

5

Hello! Thank you for the clean + user friendly codebase! I'm trying to finetune the VQ-VAE tokenizer and noticed some keys might be missing from the pretrained checkpoint listed on...

julian-q

Hello, due to limited computing resources, it is difficult for me to reproduce Llamagen from the beginning of training. If I only need to perform inference fine-tuning, what should I...

heavenhellchen

Training cost

3

Thanks for the amazing work, could you open the training cost for each model? such as training GPU times and the least GPU needed.

liming-ai