Update README.md

README.md

@@ -5,6 +5,7 @@ base_model:
 - nbeerbower/Mahou-1.2a-mistral-7B
 datasets:
 - flammenai/MahouMix-v1
+- flammenai/FlameMix-DPO-v1
 ---
 ![image/png](https://huggingface.co/flammenai/Mahou-1.0-mistral-7B/resolve/main/mahou1.png)
 
@@ -43,6 +44,64 @@ This model has been trained to use ChatML format.
 
 ### Method
 
-DPO finetuned for 6 epochs using an A100 on Google Colab.
+DPO finetuned using an A100 on Google Colab.
 
-[Fine-tune a Mistral-7b model with Direct Preference Optimization](https://towardsdatascience.com/fine-tune-a-mistral-7b-model-with-direct-preference-optimization-708042745aac) - [Maxime Labonne](https://huggingface.co/mlabonne)
+[Fine-tune a Mistral-7b model with Direct Preference Optimization](https://towardsdatascience.com/fine-tune-a-mistral-7b-model-with-direct-preference-optimization-708042745aac) - [Maxime Labonne](https://huggingface.co/mlabonne)
+
+### Configuration
+
+LoRA, model, and training settings:
+
+```python
+# LoRA configuration
+peft_config = LoraConfig(
+    r=16,
+    lora_alpha=16,
+    lora_dropout=0.05,
+    bias="none",
+    task_type="CAUSAL_LM",
+    target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']
+)
+# Model to fine-tune
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.bfloat16,
+    load_in_4bit=True
+)
+model.config.use_cache = False
+# Reference model
+ref_model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.bfloat16,
+    load_in_4bit=True
+)
+# Training arguments
+training_args = TrainingArguments(
+    per_device_train_batch_size=4,
+    gradient_accumulation_steps=4,
+    gradient_checkpointing=True,
+    learning_rate=5e-5,
+    lr_scheduler_type="cosine",
+    max_steps=2000,
+    save_strategy="no",
+    logging_steps=1,
+    output_dir=new_model,
+    optim="paged_adamw_32bit",
+    warmup_steps=100,
+    bf16=True,
+    report_to="wandb",
+)
+# Create DPO trainer
+dpo_trainer = DPOTrainer(
+    model,
+    ref_model,
+    args=training_args,
+    train_dataset=dataset,
+    tokenizer=tokenizer,
+    peft_config=peft_config,
+    beta=0.1,
+    force_use_ref_model=True
+)
+# Fine-tune model with DPO
+dpo_trainer.train()
+```