OpenPipe/lora_test

See axolotl config

axolotl version: 0.4.1

adapter: lora
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
base_model_config: meta-llama/Meta-Llama-3.1-8B-Instruct
bf16: true
dataset_processes: 8
datasets:
- path: /tmp/train.jsonl
  type:
    field_instruction: input
    field_output: output
    field_system: system
    format: '{instruction}'
    no_input_format: '{instruction}'
    system_prompt: ''
flash_attention: true
fp16: false
fsdp: []
gradient_accumulation_steps: 4
gradient_checkpointing: true
group_by_length: false
is_llama_derived_model: true
learning_rate: 0.0002
logging_steps: 1
lora_alpha: 16
lora_dropout: 0.05
lora_r: 8
lora_target_linear: true
lora_target_modules:
- gate_proj
- down_proj
- up_proj
- q_proj
- v_proj
- k_proj
- o_proj
lr_scheduler: cosine
micro_batch_size: 2
model_type: LlamaForCausalLM
num_epochs: 1
optimizer: adamw_bnb_8bit
output_dir: /models/loras2/40a0412a-574f-442e-8a35-32dd97008a01
pad_to_sequence_len: true
sample_packing: true
save_safetensors: true
save_strategy: 'no'
sequence_len: 8192
special_tokens:
  eos_token: <|eot_id|>
  pad_token: <|end_of_text|>
strict: true
tf32: false
tokenizer_type: AutoTokenizer
train_on_inputs: false
val_set_size: 0
wandb_project: OP Method
wandb_run_id: auth-12-1-24-gpt4o-relabeled-llama31
warmup_steps: 10
weight_decay: 0

models/loras2/40a0412a-574f-442e-8a35-32dd97008a01

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-8B-Instruct on the None dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 1

Training results

Framework versions

• PEFT 0.11.1
• Transformers 4.43.1
• Pytorch 2.2.2+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1