File size: 1,838 Bytes
2e763e2 4528ab7 2e763e2 4528ab7 2e763e2 4528ab7 2e763e2 4528ab7 2e763e2 4528ab7 2e763e2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 |
wandb:
entity: null
resume: auto
run_id: 33sdhxdk
experiment:
project: training
name: show-o-dpo
output_dir: train_set_hq_prune_full_t2i_dpo_10k_1e-5_v2_new_2
save_every: 10000
eval_every: 2500
generate_every: 1000
log_every: 1
log_grad_norm_every: 20
logging_dir: train_set_hq_prune_full_t2i_dpo_10k_1e-5_v2_new_2/logs
model:
rw_model: train_set_hq_prune_full
vq_model:
type: magvitv2
vq_model_name: showlab/magvitv2
showo:
load_from_showo: false
pretrained_model_path: train_set_hq_prune_full_t2i_dpo_10k_1e-5
w_clip_vit: false
vocab_size: 58498
llm_vocab_size: 50295
llm_model_path: microsoft/phi-1_5
codebook_size: 8192
num_vq_tokens: 256
num_new_special_tokens: 10
gradient_checkpointing: true
dataset:
gen_type: t2i
params:
data_path: t2i_dpo_draft.yaml
validation_prompts_file: validation_prompts/showoprompts.txt
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
optimizer:
name: adamw
params:
learning_rate: 1.0e-05
scale_lr: false
beta1: 0.9
beta2: 0.999
weight_decay: 0.01
epsilon: 1.0e-08
lr_scheduler:
scheduler: cosine
params:
learning_rate: ${optimizer.params.learning_rate}
warmup_ratio: 0.1
training:
gradient_accumulation_steps: 1
batch_size_t2i: 10
mixed_precision: bf16
enable_tf32: true
seed: 10086
overfit_one_batch: false
cond_dropout_prob: 0.1
min_masking_rate: 0.0
label_smoothing: 0.0
max_grad_norm: null
guidance_scale: 0.0
generation_timesteps: 12
beta: 0.1
reward_coef: 0
dpo_coef: 1
sft_coef: 0
num_epoch: 1
inference:
generation_timesteps: 18
guidance_scale: 1.75
|