Create axolotl-dpo.yaml (#6)

- Create axolotl-dpo.yaml (28e29a8b1e945b5cb8c27d3748d86df1673ef114)

axolotl-dpo.yaml

@@ -0,0 +1,75 @@
+base_model: meta-llama/Meta-Llama-3-8B-Instruct
+model_type: AutoModelForCausalLM
+tokenizer_type: AutoTokenizer
+
+load_in_8bit: false
+load_in_4bit: true
+strict: false
+
+save_safetensors: true
+
+rl: dpo
+chat_template: chatml
+datasets:
+  - path: Intel/orca_dpo_pairs
+    split: train
+    type: chatml.intel
+
+dataset_prepared_path:
+val_set_size: 0.1
+output_dir: ./models/Meta-Llama-3-8B-instruct-DPO-v0.3
+
+adapter: qlora
+lora_model_dir:
+
+sequence_len: 8192
+sample_packing: false
+pad_to_sequence_len: false
+
+lora_r: 64
+lora_alpha: 32
+lora_dropout: 0.05
+lora_target_linear: true
+lora_fan_in_fan_out:
+lora_modules_to_save:
+  - embed_tokens
+  - lm_head
+
+wandb_project: llama-3-instruct-dpo
+wandb_entity:
+wandb_watch:
+wandb_name:
+wandb_log_model:
+
+gradient_accumulation_steps: 2
+micro_batch_size: 1
+num_epochs: 3
+optimizer: paged_adamw_32bit
+lr_scheduler: cosine
+learning_rate: 5e-6
+train_on_inputs: false
+group_by_length: false
+
+bf16: auto
+fp16: 
+tf32: 
+
+gradient_checkpointing: true
+early_stopping_patience:
+resume_from_checkpoint:
+local_rank:
+logging_steps: 1
+xformers_attention:
+flash_attention: true
+warmup_steps: 100
+evals_per_epoch: 1
+eval_table_size:
+eval_table_max_new_tokens: 128
+saves_per_epoch: 6
+debug:
+deepspeed:
+weight_decay: 0.0
+fsdp:
+fsdp_config:
+special_tokens:
+  pad_token: "<|end_of_text|>"