| base_model: microsoft/phi-1_5 |
| base_model_config: microsoft/phi-1_5 |
| model_type: AutoModelForCausalLM |
| tokenizer_type: AutoTokenizer |
| is_llama_derived_model: false |
| trust_remote_code: true |
|
|
| load_in_8bit: false |
| load_in_4bit: false |
| strict: false |
|
|
| datasets: |
| - path: openaccess-ai-collective/oo-gpt4-filtered |
| type: alpaca_w_system.load_open_orca_chatml |
| data_files: |
| - 1M-GPT4-Augmented-filtered-gt10.parquet |
|
|
| dataset_prepared_path: last_run_prepared |
| val_set_size: 0.01 |
| output_dir: ./phi-oo-out |
| hub_model_id: Open-Orca/oo-phi-1_5 |
|
|
| sequence_len: 2048 |
| sample_packing: false |
| pad_to_sequence_len: |
|
|
| adapter: |
| lora_model_dir: |
| lora_r: |
| lora_alpha: |
| lora_dropout: |
| lora_target_linear: |
| lora_fan_in_fan_out: |
|
|
| wandb_project: phi-oo |
| wandb_entity: open-orca |
| wandb_watch: |
| wandb_run_id: |
| wandb_log_model: |
|
|
| gradient_accumulation_steps: 1 |
| micro_batch_size: 3 |
| num_epochs: 5 |
| optimizer: adamw_torch |
| adam_beta2: 0.95 |
| adam_epsilon: 0.00001 |
| max_grad_norm: 1.0 |
| lr_scheduler: cosine |
| learning_rate: 0.000003 |
|
|
| train_on_inputs: false |
| group_by_length: true |
| bf16: true |
| fp16: false |
| tf32: true |
|
|
| gradient_checkpointing: |
| early_stopping_patience: |
| resume_from_checkpoint: |
| local_rank: |
| logging_steps: 1 |
| xformers_attention: |
| flash_attention: |
|
|
| warmup_steps: 1000 |
| eval_steps: 0.02 |
| save_steps: 0.10 |
| do_bench_eval: true |
| bench_dataset: "pharaouk/dharma-1/dharma_1_full.json" |
| debug: |
| deepspeed: |
| weight_decay: 0.1 |
| fsdp: |
| fsdp_config: |
| resize_token_embeddings_to_32x: true |
| special_tokens: |
| bos_token: "<|endoftext|>" |
| eos_token: "<|im_end|>" |
| unk_token: "<|endoftext|>" |
| pad_token: "<|endoftext|>" |
| tokens: |
| - "<|im_start|>" |
| - "<|im_end|>" |
|
|
