accelerator_config:
  dispatch_batches: null
  even_batches: true
  gradient_accumulation_kwargs: null
  non_blocking: false
  split_batches: false
  use_seedable_sampler: true
activation_checkpointing: true
activation_checkpointing_kwargs: null
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1.0e-08
auto_find_batch_size: false
average_tokens_across_devices: true
batch_eval_metrics: false
beta: 0.1
bf16: true
bf16_full_eval: false
data_seed: null
dataloader_drop_last: false
dataloader_in_order: true
dataloader_multiprocessing_context: null
dataloader_num_workers: 0
dataloader_persistent_workers: false
dataloader_pin_memory: true
dataloader_prefetch_factor: null
dataset_num_proc: null
ddp_backend: null
ddp_broadcast_buffers: null
ddp_bucket_cap_mb: null
ddp_find_unused_parameters: null
ddp_static_graph: null
ddp_timeout: 1800
debug: []
deepspeed: null
disable_dropout: true
disable_tqdm: false
do_eval: false
do_predict: false
do_train: false
enable_jit_checkpoint: false
eval_accumulation_steps: null
eval_delay: 0
eval_do_concat_batches: true
eval_on_start: false
eval_steps: null
eval_strategy: 'no'
eval_use_gather_object: false
fp16: false
fp16_full_eval: false
fsdp: null
fsdp_config: null
full_determinism: false
generate_during_eval: false
gradient_accumulation_steps: 16
gradient_checkpointing: true
gradient_checkpointing_kwargs: null
greater_is_better: null
hub_always_push: false
hub_model_id: null
hub_private_repo: null
hub_revision: null
hub_strategy: every_save
hub_token: <HUB_TOKEN>
ignore_data_skip: false
include_for_metrics: []
include_num_input_tokens_seen: 'no'
is_encoder_decoder: null
label_names: null
label_smoothing_factor: 0.0
learning_rate: 1.0e-05
length_column_name: length
liger_kernel_config: null
load_best_model_at_end: false
local_rank: -1
log_level: passive
log_level_replica: warning
log_on_each_node: true
logging_first_step: false
logging_nan_inf_filter: true
logging_steps: 2
logging_strategy: steps
lr_scheduler_kwargs: null
lr_scheduler_type: linear
max_completion_length: null
max_density_ratio: 10.0
max_grad_norm: 1.0
max_length: 2176
max_steps: -1
metric_for_best_model: null
min_density_ratio: 0.5
model_init_kwargs: null
neftune_noise_alpha: null
num_train_epochs: 3.0
optim: adamw_torch_fused
optim_args: null
optim_target_modules: null
output_dir: /tmp/out/adapter
parallelism_config: null
per_device_eval_batch_size: 2
per_device_train_batch_size: 1
precompute_ref_log_probs: false
prediction_loss_only: false
project: huggingface
prompt_sample_size: 1024
push_to_hub: false
remove_unused_columns: false
report_to: []
restore_callback_states_from_checkpoint: false
resume_from_checkpoint: null
run_name: null
save_on_each_node: false
save_only_model: false
save_steps: 500
save_strategy: 'no'
save_total_limit: null
seed: 42
skip_memory_metrics: true
tf32: null
torch_compile: false
torch_compile_backend: null
torch_compile_mode: null
torch_empty_cache_steps: null
trackio_bucket_id: null
trackio_space_id: null
trackio_static_space_id: null
train_sampling_strategy: random
trust_remote_code: false
use_cache: false
use_cpu: false
use_liger_kernel: false
warmup_steps: 0
weight_decay: 0.0