See axolotl config

axolotl version: 0.9.0

base_model: meta-llama/Llama-3.1-8B-Instruct
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer
gradient_accumulation_steps: 2
micro_batch_size: 8
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0001
load_in_8bit: true
load_in_4bit: false
adapter: lora
lora_model_dir: null
lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
- q_proj
- v_proj
- k_proj
datasets:
- path: /workspace/FinLoRA/data/train/headline_train.jsonl
  type:
    system_prompt: ''
    field_system: system
    field_instruction: context
    field_output: target
    format: '[INST] {instruction} [/INST]'
    no_input_format: '[INST] {instruction} [/INST]'
dataset_prepared_path: null
val_set_size: 0.02
output_dir: /workspace/FinLoRA/lora/axolotl-output/headline_llama_3_1_8b_8bits_r8_dora
peft_use_dora: true
sequence_len: 4096
sample_packing: false
pad_to_sequence_len: false
wandb_project: finlora_models
wandb_entity: null
wandb_watch: gradients
wandb_name: headline_llama_3_1_8b_8bits_r8_dora
wandb_log_model: 'false'
bf16: auto
tf32: false
gradient_checkpointing: true
resume_from_checkpoint: null
logging_steps: 500
flash_attention: false
deepspeed: deepspeed_configs/zero1.json
warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
weight_decay: 0.0
special_tokens:
  pad_token: <|end_of_text|>
chat_template: llama3

workspace/FinLoRA/lora/axolotl-output/headline_llama_3_1_8B_8bits_r8_dora

This model is a fine-tuned version of meta-llama/Llama-3.1-8B-Instruct on the /workspace/FinLoRA/data/train/headline_train.jsonl dataset. It achieves the following results on the evaluation set:

• Loss: 0.0469

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 4.0

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0008	1	7.2540
No log	0.2503	315	0.0536
0.2177	0.5006	630	0.0576
0.2177	0.7509	945	0.0422
0.0452	1.0008	1260	0.0463
0.039	1.2511	1575	0.0439
0.039	1.5014	1890	0.0469
0.0359	1.7517	2205	0.0415
0.0324	2.0016	2520	0.0392
0.0324	2.2519	2835	0.0451
0.0273	2.5022	3150	0.0426
0.0273	2.7525	3465	0.0415
0.0279	3.0024	3780	0.0414
0.0231	3.2527	4095	0.0430
0.0231	3.5030	4410	0.0465
0.0187	3.7533	4725	0.0469

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0.dev20250319+cu128
• Datasets 3.5.0
• Tokenizers 0.21.1