Meta-Llama-3-8B-Instruct-peft-for-humor

<!-- 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: 0.13.0.dev0

adapter: qlora
base_model: NousResearch/Meta-Llama-3-8B-Instruct
bf16: auto
chat_template: llama3
dataset_prepared_path: last_run_prepared
datasets:
- path: hfht/peft-for-humor-train-dataset
  split: train
  type: chat_template
do_causal_lm_eval: false
evals_per_epoch: 1
flash_attention: true
gradient_accumulation_steps: 4
gradient_checkpointing: true
learning_rate: 0.0006
load_in_4bit: true
load_in_8bit: false
logging_steps: 1
lora_alpha: 64
lora_dropout: 0
lora_mlp_kernel: true
lora_model_dir: null
lora_modules_to_save:
- embed_tokens
- lm_head
lora_o_kernel: true
lora_qkv_kernel: true
lora_r: 64
lora_target_linear: true
lr_scheduler: cosine
micro_batch_size: 2
num_epochs: 4
optimizer: adamw_bnb_8bit
output_dir: ./out/loras/humor
resume_from_checkpoint: null
sample_packing: true
save_strategy: 'no'
save_total_limit: 1
saves_per_epoch: 0
sequence_len: 4096
special_tokens:
  pad_token: <|finetune_right_pad_id|>
tf32: false
train_on_inputs: false
val_set_size: 0.005
warmup_steps: 5
weight_decay: 0.0

</details><br>

out/loras/humor

This model is a fine-tuned version of NousResearch/Meta-Llama-3-8B-Instruct on the hfht/peft-for-humor-train-dataset dataset. It achieves the following results on the evaluation set:

• Loss: 2.9864
• Memory/max Active (gib): 18.36
• Memory/max Allocated (gib): 18.36
• Memory/device Reserved (gib): 32.49

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: 0.0006
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 5
• training_steps: 32

Training results

Framework versions

• PEFT 0.18.0
• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.4.1
• Tokenizers 0.22.1