See axolotl config

axolotl version: 0.6.0

datasets:
- path: phunguyen01/test-sft-data
  split: train
  type: chat_template
  field_messages: messages
  message_field_role: role
  message_field_content: content
  roles:
    system:
    - system
    user:
    - user
    assistant:
    - assistant
test_datasets: null
sequence_len: 2048
base_model: Qwen/Qwen2.5-0.5B-Instruct
adapter: null
lora_r: null
lora_alpha: null
lora_dropout: 0.05
lora_target_linear: true
micro_batch_size: 4
num_epochs: 3
learning_rate: 1.0e-06
output_dir: experiments/9886f198-d47e-453b-8fb7-8b50d0164301
dataset_prepared_path: experiments/9886f198-d47e-453b-8fb7-8b50d0164301/dataset_prepared
seed: 42
gradient_accumulation_steps: 1
flash_attention: true
train_on_inputs: false
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true
bf16: auto
logging_steps: 10
lr_scheduler: cosine
warmup_ratio: null
weight_decay: null
save_strategy: epoch
evals_per_epoch: 1
wandb_project: llm-training-platform
wandb_name: test-sft
wandb_run_id: 15273306-5993-4075-846c-58baf048e50e
hub_model_id: phunguyen01/quiin-sft-test
push_to_hub: true

quiin-sft-test

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B-Instruct on the phunguyen01/test-sft-data 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: 1e-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• optimizer: Use OptimizerNames.ADAMW_HF 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
• num_epochs: 3

Training results

Framework versions

• Transformers 4.46.3
• Pytorch 2.3.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3