formula_llama_3_1_8b_fp16_r8

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.9.1.post1

base_model: meta-llama/Llama-3.1-8B-Instruct
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer
gradient_accumulation_steps: 2
micro_batch_size: 4
num_epochs: 1
learning_rate: 0.0001
optimizer: adamw_torch_fused
lr_scheduler: cosine
load_in_8bit: false
load_in_4bit: false
adapter: lora
lora_r: 8
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
- q_proj
- k_proj
- v_proj
datasets:
- path: /workspace/FinLoRA/data/train/formula_train.jsonl
  type:
    field_instruction: context
    field_output: target
    format: '[INST] {instruction} [/INST]'
    no_input_format: '[INST] {instruction} [/INST]'
val_set_size: 0.02
output_dir: /workspace/FinLoRA/lora/axolotl-output/formula_llama_3_1_8b_fp16_r8
sequence_len: 4096
gradient_checkpointing: true
logging_steps: 500
warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
weight_decay: 0.0
special_tokens:
  pad_token: <|end_of_text|>
deepspeed: deepspeed_configs/zero1.json
bf16: auto
tf32: false
chat_template: llama3
wandb_name: formula_llama_3_1_8b_fp16_r8

</details><br>

workspace/FinLoRA/lora/axolotl-output/formula_llama_3_1_8b_fp16_r8

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

• Loss: 0.8450

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: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 3
• gradient_accumulation_steps: 2
• total_train_batch_size: 24
• total_eval_batch_size: 12
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 1.0

Training results

Framework versions

• PEFT 0.15.2
• Transformers 4.51.3
• Pytorch 2.8.0.dev20250319+cu128
• Datasets 3.5.1
• Tokenizers 0.21.1