metadata
dataset_info:
features:
- name: Model_name
dtype: string
- name: Train_size
dtype: int64
- name: Test_size
dtype: int64
- name: arg
struct:
- name: adafactor
dtype: bool
- name: adam_beta1
dtype: float64
- name: adam_beta2
dtype: float64
- name: adam_epsilon
dtype: float64
- name: bf16
dtype: bool
- name: fp16
dtype: bool
- name: fp16_opt_level
dtype: string
- name: gradient_accumulation_steps
dtype: int64
- name: half_precision_backend
dtype: string
- name: label_smoothing_factor
dtype: float64
- name: learning_rate
dtype: float64
- name: lr_scheduler_type
dtype: string
- name: max_grad_norm
dtype: float64
- name: max_steps
dtype: int64
- name: n_gpu
dtype: int64
- name: num_train_epochs
dtype: int64
- name: optim
dtype: string
- name: optim_args
dtype: string
- name: per_device_eval_batch_size
dtype: int64
- name: per_device_train_batch_size
dtype: int64
- name: warmup_ratio
dtype: float64
- name: warmup_steps
dtype: int64
- name: weight_decay
dtype: float64
- name: lora
sequence: string
- name: Parameters
dtype: int64
- name: Trainable_parameters
dtype: int64
- name: r
dtype: int64
- name: Memory Allocation
dtype: string
- name: Training Time
dtype: string
- name: accuracy
dtype: float64
- name: f1_macro
dtype: float64
- name: f1_weighted
dtype: float64
- name: precision
dtype: float64
- name: recall
dtype: float64
splits:
- name: train
num_bytes: 381
num_examples: 1
download_size: 15879
dataset_size: 381
configs:
- config_name: default
data_files:
- split: train
path: data/train-*