See axolotl config

axolotl version: 0.4.1

yaml

1adapter: lora 2base_model: unsloth/SmolLM-1.7B 3bf16: true 4chat_template: llama3 5dataloader_num_workers: 24 6dataset_prepared_path: null 7datasets: 8- data_files: 9 - 77c8a61dc22b9e25_train_data.json 10 ds_type: json 11 format: custom 12 path: /workspace/input_data/77c8a61dc22b9e25_train_data.json 13 type: 14 field_instruction: source 15 field_output: target 16 format: '{instruction}' 17 no_input_format: '{instruction}' 18 system_format: '{system}' 19 system_prompt: '' 20debug: null 21deepspeed: null 22device_map: auto 23do_eval: true 24early_stopping_patience: 3 25eval_batch_size: 2 26eval_max_new_tokens: 128 27eval_steps: 150 28eval_table_size: null 29evals_per_epoch: null 30flash_attention: true 31fp16: false 32fsdp: null 33fsdp_config: null 34gradient_accumulation_steps: 8 35gradient_checkpointing: false 36group_by_length: true 37hub_model_id: abaddon182/5b92094f-c6f9-4974-aa89-d6850caa7322 38hub_repo: null 39hub_strategy: checkpoint 40hub_token: null 41learning_rate: 0.0002 42load_in_4bit: false 43load_in_8bit: false 44local_rank: null 45logging_steps: 50 46lora_alpha: 64 47lora_dropout: 0.05 48lora_fan_in_fan_out: null 49lora_model_dir: null 50lora_r: 32 51lora_target_linear: true 52lr_scheduler: cosine 53max_grad_norm: 1.0 54max_steps: 3000 55micro_batch_size: 2 56mlflow_experiment_name: /tmp/77c8a61dc22b9e25_train_data.json 57model_type: AutoModelForCausalLM 58num_epochs: 10 59optim_args: 60 adam_beta1: 0.9 61 adam_beta2: 0.999 62 adam_epsilon: 1e-8 63optimizer: adamw_torch_fused 64output_dir: miner_id_24 65pad_to_sequence_len: true 66resume_from_checkpoint: null 67s2_attention: null 68sample_packing: false 69save_steps: 150 70saves_per_epoch: null 71sequence_len: 512 72strict: false 73tf32: true 74tokenizer_type: AutoTokenizer 75train_on_inputs: false 76trust_remote_code: true 77val_set_size: 0.05 78wandb_entity: null 79wandb_mode: online 80wandb_name: 116d94b2-1780-42ba-a21d-ea5d0b4896e2 81wandb_project: Gradients-On-Demand 82wandb_run: your_name 83wandb_runid: 116d94b2-1780-42ba-a21d-ea5d0b4896e2 84warmup_steps: 50 85weight_decay: 0.0 86xformers_attention: null 87

5b92094f-c6f9-4974-aa89-d6850caa7322

• Loss: 0.9359

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.999,adam_epsilon=1e-8
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 50
• training_steps: 3000

Training results

Framework versions

• PEFT 0.13.2
• Transformers 4.46.0
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1