See axolotl config

axolotl version: 0.9.0

base_model: NousResearch/Meta-Llama-3.1-8B-Instruct
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer
gradient_accumulation_steps: 8
micro_batch_size: 1
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/finer_train_batched.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/finer_llama_3_1_8b_8bits_r8
sequence_len: 4096
sample_packing: false
pad_to_sequence_len: false
wandb_project: finlora_models
wandb_entity: null
wandb_watch: gradients
wandb_name: finer_llama_3_1_8b_8bits_r8
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/fine-tune/axolotl-output/finer_llama_3_1_8B_8bits_r8

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

• Loss: 0.0331

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: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• total_eval_batch_size: 2
• 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	0.0016	1	0.5433
No log	0.2497	153	0.0520
No log	0.4995	306	0.0459
No log	0.7492	459	0.0406
0.0693	0.9990	612	0.0386
0.0693	1.2497	765	0.0396
0.0693	1.4995	918	0.0363
0.036	1.7492	1071	0.0351
0.036	1.9990	1224	0.0348
0.036	2.2497	1377	0.0360
0.0302	2.4995	1530	0.0321
0.0302	2.7492	1683	0.0347
0.0302	2.9990	1836	0.0324
0.0302	3.2497	1989	0.0328
0.0242	3.4995	2142	0.0334
0.0242	3.7492	2295	0.0332
0.0242	3.9990	2448	0.0331

Framework versions

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