architecture:
    backbone_dtype: bfloat16
    force_embedding_gradients: false
    gradient_checkpointing: true
    intermediate_dropout: 0.0
    pretrained: true
    pretrained_weights: ''
augmentation:
    neftune_noise_alpha: 0.0
    random_parent_probability: 0.0
    skip_parent_probability: 0.0
    token_mask_probability: 0.0
dataset:
    add_eos_token_to_answer: true
    add_eos_token_to_prompt: true
    add_eos_token_to_system: true
    answer_column: label
    answer_column_label: category
    chatbot_author: ''
    chatbot_name: ''
    data_sample: 1.0
    data_sample_choice:
    - Train
    - Validation
    limit_chained_samples: false
    mask_prompt_labels: true
    num_classes: 10
    parent_id_column: None
    personalize: false
    prompt_column:
    - Description
    system_column: None
    text_answer_separator: ''
    text_prompt_start: ''
    text_system_start: ''
    train_dataframe: /app/train_df.csv
    validation_dataframe: ''
    validation_size: 0.2
    validation_strategy: automatic
environment:
    compile_model: false
    deepspeed_offload_optimizer: false
    deepspeed_reduce_bucket_size: 10000000.0
    deepspeed_stage3_max_live_parameters: 10000000.0
    deepspeed_stage3_max_reuse_distance: 10000000.0
    deepspeed_stage3_param_persistence_threshold: 10000000.0
    deepspeed_stage3_prefetch_bucket_size: 10000000.0
    find_unused_parameters: false
    gpus:
    - '0'
    huggingface_branch: main
    mixed_precision: false
    number_of_workers: 8
    seed: -1
    trust_remote_code: false
    use_deepspeed: false
experiment_name: Falcon 1B Classification
llm_backbone: tiiuae/falcon-rw-1b
logging:
    logger: None
    neptune_project: ''
output_directory: /app/output
prediction:
    batch_size_inference: 0
    metric: AUC
problem_type: text_causal_classification_modeling
tokenizer:
    add_prefix_space: false
    add_prompt_answer_tokens: false
    max_length: 512
    max_length_answer: 256
    max_length_prompt: 256
    padding_quantile: 1.0
    use_fast: true
tracking_mode: during_epoch
training:
    alpha: 0.9900000095
    batch_size: 16
    beta1: 0.8999999762
    beta2: 0.9990000129
    differential_learning_rate: 1.0e-05
    differential_learning_rate_layers: []
    drop_last_batch: true
    epochs: 1
    eps: 1.0e-08
    evaluate_before_training: false
    evaluation_epochs: 1.0
    grad_accumulation: 1
    gradient_clip: 0.5
    learning_rate: 0.0001
    lora: true
    lora_alpha: 16
    lora_dropout: 0.0500000007
    lora_r: 4
    lora_target_modules: []
    loss_function: CrossEntropyLoss
    momentum: 0.0
    nesterov: false
    optimizer: AdamW
    rho: 0.8999999762
    save_best_checkpoint: false
    schedule: Cosine
    train_validation_data: false
    use_flash_attention_2: false
    warmup_epochs: 0.0
    weight_decay: 0.0