| base_model: huggyllama/llama-13b | |
| base_model_config: huggyllama/llama-13b | |
| model_type: LlamaForCausalLM | |
| tokenizer_type: LlamaTokenizer | |
| load_in_8bit: false | |
| load_in_4bit: false | |
| strict: false | |
| push_dataset_to_hub: openaccess-ai-collective | |
| hf_use_auth_token: true | |
| datasets: | |
| - path: Open-Orca/oo-gpt4-200k | |
| type: alpaca_w_system.load_open_orca | |
| dataset_prepared_path: last_run_prepared | |
| val_set_size: 0.05 | |
| adapter: | |
| lora_model_dir: | |
| sequence_len: 2048 | |
| max_packed_sequence_len: 2048 | |
| lora_r: | |
| lora_alpha: | |
| lora_dropout: | |
| lora_target_modules: | |
| lora_target_linear: | |
| lora_fan_in_fan_out: | |
| wandb_project: open-orca-13b | |
| wandb_watch: | |
| wandb_run_id: | |
| wandb_log_model: | |
| output_dir: ./open-orca-13b | |
| gradient_accumulation_steps: 1 | |
| micro_batch_size: 8 | |
| num_epochs: 5 | |
| optimizer: adamw_torch | |
| adam_beta2: 0.99 | |
| torchdistx_path: | |
| lr_scheduler: cosine | |
| lr_quadratic_warmup: true | |
| learning_rate: 0.0001 | |
| train_on_inputs: false | |
| group_by_length: true | |
| bf16: true | |
| fp16: false | |
| tf32: true | |
| gradient_checkpointing: true | |
| early_stopping_patience: | |
| resume_from_checkpoint: | |
| local_rank: | |
| logging_steps: 1 | |
| xformers_attention: true | |
| flash_attention: | |
| gptq_groupsize: | |
| gptq_model_v1: | |
| warmup_steps: 100 | |
| eval_steps: 101 | |
| save_steps: 711 | |
| debug: | |
| deepspeed: | |
| weight_decay: 0.1 | |
| fsdp: | |
| - full_shard | |
| - auto_wrap | |
| fsdp_config: | |
| fsdp_offload_params: true | |
| fsdp_state_dict_type: FULL_STATE_DICT | |
| fsdp_transformer_layer_cls_to_wrap: LlamaDecoderLayer | |
| special_tokens: | |
| bos_token: "<s>" | |
| eos_token: "</s>" | |
| unk_token: "<unk>" | |