See axolotl config

axolotl version: 0.4.1

yaml

1adapter: lora 2base_model: samoline/5cf4792e-00af-4d5f-b922-4c7d5237f306 3bf16: auto 4chat_template: llama3 5dataloader_num_workers: 8 6dataset_prepared_path: null 7datasets: 8- data_files: 9 - 4b98334ce710802c_train_data.json 10 ds_type: json 11 format: custom 12 path: /workspace/input_data/ 13 type: 14 field_input: input 15 field_instruction: instruct 16 field_output: output 17 format: '{instruction} {input}' 18 no_input_format: '{instruction}' 19 system_format: '{system}' 20 system_prompt: '' 21debug: null 22deepspeed: null 23early_stopping_patience: null 24eval_max_new_tokens: 128 25eval_table_size: null 26evals_per_epoch: 4 27flash_attention: true 28fp16: null 29fsdp: null 30fsdp_config: null 31gradient_accumulation_steps: 1 32gradient_checkpointing: true 33group_by_length: false 34hub_model_id: cimol/94bee815-bb95-47e6-835f-ad38135a2860 35hub_repo: null 36hub_strategy: end 37hub_token: null 38learning_rate: 0.0001 39load_in_4bit: false 40load_in_8bit: false 41local_rank: null 42logging_steps: 1 43lora_alpha: 32 44lora_dropout: 0.15 45lora_fan_in_fan_out: null 46lora_model_dir: null 47lora_r: 32 48lora_target_linear: true 49loraplus_lr_embedding: 1.0e-06 50loraplus_lr_ratio: 16 51lr_scheduler: cosine 52max_grad_norm: 1 53max_steps: 100 54micro_batch_size: 8 55mlflow_experiment_name: /tmp/4b98334ce710802c_train_data.json 56model_type: AutoModelForCausalLM 57num_epochs: 20 58optim_args: 59 adam_beta1: 0.9 60 adam_beta2: 0.999 61 adam_epsilon: 1e-8 62optimizer: adamw_torch_fused 63output_dir: miner_id_24 64pad_to_sequence_len: true 65resume_from_checkpoint: null 66s2_attention: null 67sample_packing: false 68saves_per_epoch: 5 69sequence_len: 1024 70strict: false 71tf32: true 72tokenizer_type: AutoTokenizer 73train_on_inputs: false 74trust_remote_code: true 75val_set_size: 0.05 76wandb_entity: null 77wandb_mode: online 78wandb_name: 859088c3-5caa-40f5-84fd-490d89cece7b 79wandb_project: Gradients-On-Demand 80wandb_run: your_name 81wandb_runid: 859088c3-5caa-40f5-84fd-490d89cece7b 82warmup_steps: 150 83weight_decay: 0.0 84xformers_attention: null 85

94bee815-bb95-47e6-835f-ad38135a2860

• Loss: 0.9920

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• 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: 150
• training_steps: 100

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