|
accum_freq: 1 |
|
aug_cfg: {} |
|
batch_size: 512 |
|
beta1: 0.9 |
|
beta2: 0.98 |
|
checkpoint_path: ./logs/2024_02_13-14_16_04-model_ViT-B-32-lr_5e-05-b_512-j_8-p_amp_bf16/checkpoints |
|
coca_caption_loss_weight: 2.0 |
|
coca_contrastive_loss_weight: 1.0 |
|
copy_codebase: False |
|
csv_caption_key: captions |
|
csv_img_key: images |
|
csv_separator: |
|
dataset_resampled: False |
|
dataset_type: auto |
|
ddp_static_graph: True |
|
debug: False |
|
delete_previous_checkpoint: False |
|
device: cuda:0 |
|
dist_backend: nccl |
|
dist_url: env:// |
|
distill: False |
|
distill_model: None |
|
distill_pretrained: None |
|
distributed: False |
|
epochs: 15 |
|
epochs_cooldown: None |
|
eps: 1e-06 |
|
force_custom_text: False |
|
force_image_size: None |
|
force_patch_dropout: None |
|
force_quick_gelu: False |
|
gather_with_grad: True |
|
grad_checkpointing: False |
|
grad_clip_norm: None |
|
horovod: False |
|
image_interpolation: None |
|
image_mean: None |
|
image_resize_mode: None |
|
image_std: None |
|
imagenet_v2: None |
|
imagenet_val: None |
|
local_loss: True |
|
local_rank: 0 |
|
lock_image: False |
|
lock_image_freeze_bn_stats: False |
|
lock_image_unlocked_groups: 0 |
|
lock_text: False |
|
lock_text_freeze_layer_norm: False |
|
lock_text_unlocked_layers: 0 |
|
log_every_n_steps: 100 |
|
log_level: 20 |
|
log_local: False |
|
log_path: ./logs/2024_02_13-14_16_04-model_ViT-B-32-lr_5e-05-b_512-j_8-p_amp_bf16/out.log |
|
logs: ./logs/ |
|
lr: 5e-05 |
|
lr_cooldown_end: 0.0 |
|
lr_cooldown_power: 1.0 |
|
lr_scheduler: cosine |
|
model: ViT-B-32 |
|
name: 2024_02_13-14_16_04-model_ViT-B-32-lr_5e-05-b_512-j_8-p_amp_bf16 |
|
no_set_device_rank: False |
|
precision: amp_bf16 |
|
pretrained: laion2b_s34b_b79k |
|
pretrained_image: False |
|
rank: 0 |
|
remote_sync: None |
|
remote_sync_frequency: 300 |
|
remote_sync_protocol: s3 |
|
report_to: |
|
resume: None |
|
save_frequency: 5 |
|
save_most_recent: False |
|
seed: 0 |
|
siglip: False |
|
skip_scheduler: False |
|
tensorboard: False |
|
tensorboard_path: |
|
torchcompile: False |
|
torchscript: False |
|
trace: False |
|
train_data: /staging/jzhang2427/flickr30k_entities/train_data/train_data_pos_neg_clip.csv |
|
train_data_upsampling_factors: None |
|
train_num_samples: None |
|
use_bn_sync: False |
|
use_bnb_linear: None |
|
val_data: None |
|
val_frequency: 5 |
|
val_num_samples: None |
|
wandb: False |
|
wandb_notes: |
|
wandb_project_name: open-clip |
|
warmup: 1024 |
|
wd: 0.2 |
|
workers: 8 |
|
world_size: 1 |
|
zeroshot_frequency: 5 |
|
|