See axolotl config

axolotl version: 0.4.1

yaml

1adapter: qlora 2auto_resume_from_checkpoints: true 3base_model: Vikhrmodels/Vikhr-7B-instruct_0.4 4bf16: auto 5chat_template: llama3 6dataloader_num_workers: 12 7dataset_prepared_path: null 8datasets: 9- data_files: 10 - eab93cbd8a081369_train_data.json 11 ds_type: json 12 format: custom 13 path: /workspace/input_data/eab93cbd8a081369_train_data.json 14 type: 15 field_input: info 16 field_instruction: classical 17 field_output: modern 18 format: '{instruction} {input}' 19 no_input_format: '{instruction}' 20 system_format: '{system}' 21 system_prompt: '' 22debug: null 23deepspeed: null 24early_stopping_patience: 3 25eval_max_new_tokens: 128 26eval_steps: 50 27eval_table_size: null 28evals_per_epoch: null 29flash_attention: true 30fp16: false 31fsdp: null 32fsdp_config: null 33gradient_accumulation_steps: 16 34gradient_checkpointing: true 35group_by_length: false 36hub_model_id: error577/9b376dea-cf02-48e4-9582-5c0c293e48f1 37hub_repo: null 38hub_strategy: checkpoint 39hub_token: null 40learning_rate: 0.0002 41load_in_4bit: true 42load_in_8bit: false 43local_rank: null 44logging_steps: 1 45lora_alpha: 128 46lora_dropout: 0.05 47lora_fan_in_fan_out: null 48lora_model_dir: null 49lora_r: 64 50lora_target_linear: true 51lr_scheduler: cosine 52max_grad_norm: 1.0 53max_steps: null 54micro_batch_size: 1 55mlflow_experiment_name: /tmp/eab93cbd8a081369_train_data.json 56model_type: AutoModelForCausalLM 57num_epochs: 3 58optimizer: adamw_bnb_8bit 59output_dir: miner_id_24 60pad_to_sequence_len: true 61resume_from_checkpoint: null 62s2_attention: null 63sample_packing: false 64save_steps: 50 65sequence_len: 512 66strict: false 67tf32: false 68tokenizer_type: AutoTokenizer 69train_on_inputs: false 70trust_remote_code: true 71val_set_size: 0.0005 72wandb_entity: null 73wandb_mode: online 74wandb_name: a8f4f147-a1f8-43a0-8042-6b029c99bc63 75wandb_project: Gradients-On-Demand 76wandb_run: your_name 77wandb_runid: a8f4f147-a1f8-43a0-8042-6b029c99bc63 78warmup_steps: 100 79weight_decay: 0.01 80xformers_attention: null 81

9b376dea-cf02-48e4-9582-5c0c293e48f1

• Loss: 1.9280

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

• learning_rate: 0.0002
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• 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: 100
• num_epochs: 3

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