See axolotl config

axolotl version: 0.4.1

yaml

1adapter: lora 2base_model: unsloth/Qwen2-1.5B 3bf16: true 4chat_template: llama3 5dataset_prepared_path: null 6datasets: 7- data_files: 8 - ca89cd0a8d4278b1_train_data.json 9 ds_type: json 10 format: custom 11 path: /workspace/input_data/ca89cd0a8d4278b1_train_data.json 12 type: 13 field_input: negative 14 field_instruction: positive 15 field_output: anchor 16 format: '{instruction} {input}' 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: 2 25eval_batch_size: 4 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: 4 35gradient_checkpointing: true 36group_by_length: true 37hub_model_id: nttx/0cba31c0-34df-4626-a8af-573d7de4bc86 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: 128 47lora_dropout: 0.3 48lora_fan_in_fan_out: null 49lora_model_dir: null 50lora_r: 64 51lora_target_linear: true 52lr_scheduler: cosine 53max_grad_norm: 1.0 54max_memory: 55 0: 75GB 56max_steps: 3000 57micro_batch_size: 4 58mlflow_experiment_name: /tmp/ca89cd0a8d4278b1_train_data.json 59model_type: AutoModelForCausalLM 60num_epochs: 10 61optim_args: 62 adam_beta1: 0.9 63 adam_beta2: 0.95 64 adam_epsilon: 1e-5 65optimizer: adamw_bnb_8bit 66output_dir: miner_id_24 67pad_to_sequence_len: true 68resume_from_checkpoint: null 69s2_attention: null 70sample_packing: false 71save_steps: 150 72saves_per_epoch: null 73sequence_len: 512 74strict: false 75tf32: true 76tokenizer_type: AutoTokenizer 77train_on_inputs: false 78trust_remote_code: true 79val_set_size: 0.05 80wandb_entity: null 81wandb_mode: online 82wandb_name: 668f674e-b67a-410c-992f-3c4d986cac17 83wandb_project: Gradients-On-Demand 84wandb_run: your_name 85wandb_runid: 668f674e-b67a-410c-992f-3c4d986cac17 86warmup_steps: 50 87weight_decay: 0.0 88xformers_attention: null 89

0cba31c0-34df-4626-a8af-573d7de4bc86

• Loss: 1.5788

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 0.0002
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=adam_beta1=0.9,adam_beta2=0.95,adam_epsilon=1e-5
• 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