architecture:
    backbone_dtype: float16
    force_embedding_gradients: false
    gradient_checkpointing: false
    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: output
    chatbot_author: ''
    chatbot_name: ''
    data_sample: 1.0
    data_sample_choice: []
    limit_chained_samples: false
    mask_prompt_labels: false
    parent_id_column: None
    personalize: false
    prompt_column:
    - instruction
    system_column: None
    text_answer_separator: ''
    text_prompt_start: ''
    text_system_start: <|system|>
    train_dataframe: /app/train_df.csv
    validation_dataframe: ''
    validation_size: 0.01
    validation_strategy: automatic
environment:
    compile_model: false
    deepspeed_reduce_bucket_size: 10000000
    deepspeed_stage3_param_persistence_threshold: 10000000
    deepspeed_stage3_prefetch_bucket_size: 10000000
    find_unused_parameters: false
    gpus:
    - '0'
    huggingface_branch: main
    mixed_precision: false
    number_of_workers: 1
    seed: -1
    trust_remote_code: true
    use_deepspeed: false
experiment_name: tinyllama_causal_language
llm_backbone: TinyLlama/TinyLlama-1.1B-Chat-v0.6
logging:
    logger: None
    neptune_project: ''
output_directory: /app/output
prediction:
    batch_size_inference: 0
    do_sample: false
    max_length_inference: 256
    metric: BLEU
    metric_gpt_model: gpt-3.5-turbo-0301
    metric_gpt_template: general
    min_length_inference: 2
    num_beams: 1
    num_history: 4
    repetition_penalty: 1.2000000477
    stop_tokens: ''
    temperature: 0.3
    top_k: 0
    top_p: 1.0
problem_type: text_causal_language_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
training:
    batch_size: 2
    differential_learning_rate: 9.999999747378752e-06
    differential_learning_rate_layers: ''
    drop_last_batch: true
    epochs: 15
    evaluate_before_training: false
    evaluation_epochs: 1.0
    grad_accumulation: 1
    gradient_clip: 0.0
    learning_rate: 0.00015
    lora: true
    lora_alpha: 16
    lora_dropout: 0.0500000007
    lora_r: 4
    lora_target_modules: ''
    loss_function: SampleAveragedCrossEntropy
    optimizer: AdamW
    save_best_checkpoint: false
    schedule: Cosine
    train_validation_data: false
    use_flash_attention_2: false
    warmup_epochs: 0.0
    weight_decay: 0.0