|
wandb: |
|
entity: null |
|
|
|
resume: 'auto' |
|
|
|
experiment: |
|
project: "demo" |
|
name: "show-o-demo" |
|
output_dir: "show-o-demo" |
|
|
|
model: |
|
vq_model: |
|
type: "magvitv2" |
|
vq_model_name: "showlab/magvitv2" |
|
|
|
showo: |
|
pretrained_model_path: "showlab/show-o-w-clip-vit" |
|
w_clip_vit: True |
|
vocab_size: 58498 |
|
llm_vocab_size: 50295 |
|
llm_model_path: 'microsoft/phi-1_5' |
|
codebook_size: 8192 |
|
num_vq_tokens: 256 |
|
|
|
gradient_checkpointing: True |
|
enable_xformers_memory_efficient_attention: True |
|
|
|
|
|
dataset: |
|
gen_type: "t2i" |
|
und_type: "large_cap" |
|
params: |
|
batch_size: ${training.batch_size} |
|
shuffle_buffer_size: 1000 |
|
num_workers: 32 |
|
resolution: 256 |
|
pin_memory: True |
|
persistent_workers: True |
|
|
|
preprocessing: |
|
max_seq_length: 128 |
|
resolution: 256 |
|
center_crop: False |
|
random_flip: False |
|
|
|
training: |
|
gradient_accumulation_steps: 1 |
|
cond_dropout_prob: 0.1 |
|
batch_size: 20 |
|
|
