wandb/run-20240720_090034-r2hlwirh/files/config.yaml

wandb_version: 1

_wandb:
 desc: null
 value:
 python_version: 3.10.13
 cli_version: 0.17.4
 framework: huggingface
 huggingface_version: 4.42.3
 is_jupyter_run: true
 is_kaggle_kernel: true
 start_time: 1721466034
 t:
 1:
 - 1
 - 2
 - 3
 - 5
 - 11
 - 12
 - 41
 - 49
 - 51
 - 53
 - 55
 - 71
 - 98
 - 105
 2:
 - 1
 - 2
 - 3
 - 5
 - 11
 - 12
 - 41
 - 49
 - 51
 - 53
 - 55
 - 71
 - 98
 - 105
 3:
 - 7
 - 23
 - 62
 - 66
 4: 3.10.13
 5: 0.17.4
 6: 4.42.3
 8:
 - 1
 - 2
 - 5
 9:
 1: transformers_trainer
 13: linux-x86_64
 m:
 - 1: train/global_step
 6:
 - 3
 - 1: train/loss
 5: 1
 6:
 - 1
 - 1: train/grad_norm
 5: 1
 6:
 - 1
 - 1: train/learning_rate
 5: 1
 6:
 - 1
 - 1: train/epoch
 5: 1
 6:
 - 1
output_dir:
 desc: null
 value: ./
overwrite_output_dir:
 desc: null
 value: false
do_train:
 desc: null
 value: false
do_eval:
 desc: null
 value: false
do_predict:
 desc: null
 value: false
eval_strategy:
 desc: null
 value: 'no'
prediction_loss_only:
 desc: null
 value: false
per_device_train_batch_size:
 desc: null
 value: 1
per_device_eval_batch_size:
 desc: null
 value: 8
per_gpu_train_batch_size:
 desc: null
 value: null
per_gpu_eval_batch_size:
 desc: null
 value: null
gradient_accumulation_steps:
 desc: null
 value: 1
eval_accumulation_steps:
 desc: null
 value: null
eval_delay:
 desc: null
 value: 0
learning_rate:
 desc: null
 value: 1.0e-05
weight_decay:
 desc: null
 value: 0.01
adam_beta1:
 desc: null
 value: 0.9
adam_beta2:
 desc: null
 value: 0.999
adam_epsilon:
 desc: null
 value: 1.0e-08
max_grad_norm:
 desc: null
 value: 1.0
num_train_epochs:
 desc: null
 value: 1
max_steps:
 desc: null
 value: -1
lr_scheduler_type:
 desc: null
 value: linear
lr_scheduler_kwargs:
 desc: null
 value: {}
warmup_ratio:
 desc: null
 value: 0.0
warmup_steps:
 desc: null
 value: 0
log_level:
 desc: null
 value: passive
log_level_replica:
 desc: null
 value: warning
log_on_each_node:
 desc: null
 value: true
logging_dir:
 desc: null
 value: ./runs/Jul20_09-00-51_caa283505014
logging_strategy:
 desc: null
 value: steps
logging_first_step:
 desc: null
 value: false
logging_steps:
 desc: null
 value: 1
logging_nan_inf_filter:
 desc: null
 value: true
save_strategy:
 desc: null
 value: steps
save_steps:
 desc: null
 value: 500
save_total_limit:
 desc: null
 value: null
save_safetensors:
 desc: null
 value: true
save_on_each_node:
 desc: null
 value: false
save_only_model:
 desc: null
 value: false
restore_callback_states_from_checkpoint:
 desc: null
 value: false
no_cuda:
 desc: null
 value: false
use_cpu:
 desc: null
 value: false
use_mps_device:
 desc: null
 value: false
seed:
 desc: null
 value: 42
data_seed:
 desc: null
 value: null
jit_mode_eval:
 desc: null
 value: false
use_ipex:
 desc: null
 value: false
bf16:
 desc: null
 value: false
fp16:
 desc: null
 value: false
fp16_opt_level:
 desc: null
 value: O1
half_precision_backend:
 desc: null
 value: auto
bf16_full_eval:
 desc: null
 value: false
fp16_full_eval:
 desc: null
 value: false
tf32:
 desc: null
 value: null
local_rank:
 desc: null
 value: 0
ddp_backend:
 desc: null
 value: null
tpu_num_cores:
 desc: null
 value: null
tpu_metrics_debug:
 desc: null
 value: false
debug:
 desc: null
 value: []
dataloader_drop_last:
 desc: null
 value: false
eval_steps:
 desc: null
 value: null
dataloader_num_workers:
 desc: null
 value: 0
dataloader_prefetch_factor:
 desc: null
 value: null
past_index:
 desc: null
 value: -1
run_name:
 desc: null
 value: ./
disable_tqdm:
 desc: null
 value: false
remove_unused_columns:
 desc: null
 value: false
label_names:
 desc: null
 value: null
load_best_model_at_end:
 desc: null
 value: false
metric_for_best_model:
 desc: null
 value: null
greater_is_better:
 desc: null
 value: null
ignore_data_skip:
 desc: null
 value: false
fsdp:
 desc: null
 value: []
fsdp_min_num_params:
 desc: null
 value: 0
fsdp_config:
 desc: null
 value:
 min_num_params: 0
 xla: false
 xla_fsdp_v2: false
 xla_fsdp_grad_ckpt: false
fsdp_transformer_layer_cls_to_wrap:
 desc: null
 value: null
accelerator_config:
 desc: null
 value:
 split_batches: false
 dispatch_batches: null
 even_batches: true
 use_seedable_sampler: true
 non_blocking: false
 gradient_accumulation_kwargs: null
deepspeed:
 desc: null
 value: null
label_smoothing_factor:
 desc: null
 value: 0.0
optim:
 desc: null
 value: adamw_torch
optim_args:
 desc: null
 value: null
adafactor:
 desc: null
 value: false
group_by_length:
 desc: null
 value: false
length_column_name:
 desc: null
 value: length
report_to:
 desc: null
 value:
 - wandb
ddp_find_unused_parameters:
 desc: null
 value: null
ddp_bucket_cap_mb:
 desc: null
 value: null
ddp_broadcast_buffers:
 desc: null
 value: null
dataloader_pin_memory:
 desc: null
 value: true
dataloader_persistent_workers:
 desc: null
 value: false
skip_memory_metrics:
 desc: null
 value: true
use_legacy_prediction_loop:
 desc: null
 value: false
push_to_hub:
 desc: null
 value: true
resume_from_checkpoint:
 desc: null
 value: null
hub_model_id:
 desc: null
 value: VQA2
hub_strategy:
 desc: null
 value: every_save
hub_token:
 desc: null
 value: <HUB_TOKEN>
hub_private_repo:
 desc: null
 value: false
hub_always_push:
 desc: null
 value: false
gradient_checkpointing:
 desc: null
 value: false
gradient_checkpointing_kwargs:
 desc: null
 value: null
include_inputs_for_metrics:
 desc: null
 value: false
eval_do_concat_batches:
 desc: null
 value: true
fp16_backend:
 desc: null
 value: auto
evaluation_strategy:
 desc: null
 value: null
push_to_hub_model_id:
 desc: null
 value: null
push_to_hub_organization:
 desc: null
 value: null
push_to_hub_token:
 desc: null
 value: <PUSH_TO_HUB_TOKEN>
mp_parameters:
 desc: null
 value: ''
auto_find_batch_size:
 desc: null
 value: false
full_determinism:
 desc: null
 value: false
torchdynamo:
 desc: null
 value: null
ray_scope:
 desc: null
 value: last
ddp_timeout:
 desc: null
 value: 1800
torch_compile:
 desc: null
 value: false
torch_compile_backend:
 desc: null
 value: null
torch_compile_mode:
 desc: null
 value: null
dispatch_batches:
 desc: null
 value: null
split_batches:
 desc: null
 value: null
include_tokens_per_second:
 desc: null
 value: false
include_num_input_tokens_seen:
 desc: null
 value: false
neftune_noise_alpha:
 desc: null
 value: null
optim_target_modules:
 desc: null
 value: null
batch_eval_metrics:
 desc: null
 value: false
eval_on_start:
 desc: null
 value: false