mirror of
https://github.com/Vision-CAIR/MiniGPT-4.git
synced 2025-04-18 03:30:50 +00:00
57 lines
997 B
YAML
57 lines
997 B
YAML
model:
|
|
arch: bind_gpt4
|
|
model_type: pretrain_vicuna
|
|
freeze_imagebind: True
|
|
freeze_qformer: False
|
|
|
|
|
|
datasets:
|
|
laion:
|
|
vis_processor:
|
|
train:
|
|
name: "imagebind_vision_train"
|
|
image_size: 224
|
|
text_processor:
|
|
train:
|
|
name: "imagebind_caption"
|
|
sample_ratio: 115
|
|
cc_sbu:
|
|
vis_processor:
|
|
train:
|
|
name: "imagebind_vision_train"
|
|
image_size: 224
|
|
text_processor:
|
|
train:
|
|
name: "imagebind_caption"
|
|
sample_ratio: 14
|
|
|
|
|
|
run:
|
|
task: imagebind_qformer_train
|
|
# optimizer
|
|
lr_sched: "linear_warmup_cosine_lr"
|
|
init_lr: 1e-4
|
|
min_lr: 8e-5
|
|
warmup_lr: 1e-6
|
|
|
|
weight_decay: 0.05
|
|
max_epoch: 4
|
|
batch_size_train: 64
|
|
batch_size_eval: 64
|
|
num_workers: 4
|
|
warmup_steps: 5000
|
|
iters_per_epoch: 5000
|
|
|
|
seed: 42
|
|
output_dir: "output/bindgpt4"
|
|
|
|
amp: True
|
|
resume_ckpt_path: null
|
|
|
|
evaluate: False
|
|
train_splits: ["train"]
|
|
|
|
device: "cuda"
|
|
world_size: 1
|
|
dist_url: "env://"
|
|
distributed: True |