|
model: |
|
arch: mini_gpt4_llama_v2 |
|
model_type: pretrain_vicuna |
|
freeze_vit: True |
|
freeze_qformer: True |
|
max_txt_len: 512 |
|
low_resource: True |
|
image_size: 224 |
|
end_sym: "</s>" |
|
llama_model: "mistralai/Mistral-7B-Instruct-v0.2" |
|
ckpt: "checkpoints/video_mistral_all_checkpoint_last.pth" |
|
use_grad_checkpoint: True |
|
chat_template: True |
|
lora_r: 64 |
|
lora_alpha: 16 |
|
length: 50 |
|
use_grad_checkpoint_llm: True |
|
max_context_len: 7200 |
|
|
|
|
|
datasets: |
|
video_chatgpt: |
|
batch_size: 1 |
|
vis_processor: |
|
train: |
|
name: "blip2_image_train" |
|
image_size: 224 |
|
text_processor: |
|
train: |
|
name: "blip_caption" |
|
sample_ratio: 200 |
|
|
|
|
|
run: |
|
task: image_text_pretrain |
|
seed: 42 |
|
amp: True |
