setup FinetuneConfig

python/prompt_bench/finetune.py

@@ -14,9 +14,11 @@ from __future__ import annotations
 
 import json
 import logging
+from dataclasses import dataclass
 from pathlib import Path
 from typing import Annotated
 
+import tomllib
 import typer
 from unsloth import FastLanguageModel
 from datasets import Dataset
@@ -25,32 +27,49 @@ from trl import SFTTrainer
 
 logger = logging.getLogger(__name__)
 
-BASE_MODEL = "unsloth/Qwen3-4B-Base-unsloth-bnb-4bit"
 
-# LoRA hyperparameters
+@dataclass
-LORA_RANK = 32
+class LoraConfig:
-LORA_ALPHA = 32
+    """LoRA adapter hyperparameters."""
-LORA_DROPOUT = 0.0
-LORA_TARGETS = [
-    "q_proj",
-    "k_proj",
-    "v_proj",
-    "o_proj",
-    "gate_proj",
-    "up_proj",
-    "down_proj",
-]
 
-# Training hyperparameters tuned for ~2k examples on a 3090
+    rank: int
-LEARNING_RATE = 2e-4
+    alpha: int
-EPOCHS = 3
+    dropout: float
-BATCH_SIZE = 2
+    targets: list[str]
-GRADIENT_ACCUMULATION = 8  # effective batch = 16
+
-MAX_SEQ_LENGTH = 4096
+
-WARMUP_RATIO = 0.05
+@dataclass
-WEIGHT_DECAY = 0.01
+class TrainingConfig:
-LOGGING_STEPS = 10
+    """Training loop hyperparameters."""
-SAVE_STEPS = 100
+
+    learning_rate: float
+    epochs: int
+    batch_size: int
+    gradient_accumulation: int
+    max_seq_length: int
+    warmup_ratio: float
+    weight_decay: float
+    logging_steps: int
+    save_steps: int
+
+
+@dataclass
+class FinetuneConfig:
+    """Top-level finetune configuration."""
+
+    base_model: str
+    lora: LoraConfig
+    training: TrainingConfig
+
+    @classmethod
+    def from_toml(cls, config_path: Path) -> FinetuneConfig:
+        """Load finetune config from a TOML file."""
+        raw = tomllib.loads(config_path.read_text())["finetune"]
+        return cls(
+            base_model=raw["base_model"],
+            lora=LoraConfig(**raw["lora"]),
+            training=TrainingConfig(**raw["training"]),
+        )
 
 
 def _messages_to_chatml(messages: list[dict]) -> str:
@@ -95,12 +114,10 @@ def main(
     output_dir: Annotated[Path, typer.Option("--output-dir", help="Where to save the merged model")] = Path(
         "output/qwen-bill-summarizer",
     ),
-    base_model: Annotated[str, typer.Option("--base-model", help="Unsloth model ID")] = BASE_MODEL,
+    config_path: Annotated[
-    epochs: Annotated[int, typer.Option("--epochs", help="Training epochs")] = EPOCHS,
+        Path,
-    batch_size: Annotated[int, typer.Option("--batch-size", help="Per-device batch size")] = BATCH_SIZE,
+        typer.Option("--config", help="TOML config file"),
-    learning_rate: Annotated[float, typer.Option("--lr", help="Learning rate")] = LEARNING_RATE,
+    ] = Path(__file__).parent / "config.toml",
-    lora_rank: Annotated[int, typer.Option("--lora-rank", help="LoRA rank")] = LORA_RANK,
-    max_seq_length: Annotated[int, typer.Option("--max-seq-length", help="Max sequence length")] = MAX_SEQ_LENGTH,
     save_gguf: Annotated[bool, typer.Option("--save-gguf/--no-save-gguf", help="Also save GGUF")] = False,
 ) -> None:
     """Fine-tune Qwen 3.5 4B on bill summarization with Unsloth + QLoRA."""
@@ -110,21 +127,23 @@ def main(
         message = f"Dataset not found: {dataset_path}"
         raise typer.BadParameter(message)
 
-    logger.info("Loading base model: %s", base_model)
+    config = FinetuneConfig.from_toml(config_path)
+
+    logger.info("Loading base model: %s", config.base_model)
     model, tokenizer = FastLanguageModel.from_pretrained(
-        model_name=base_model,
+        model_name=config.base_model,
-        max_seq_length=max_seq_length,
+        max_seq_length=config.training.max_seq_length,
         load_in_4bit=True,
         dtype=None,
     )
 
-    logger.info("Applying LoRA (rank=%d, alpha=%d)", lora_rank, LORA_ALPHA)
+    logger.info("Applying LoRA (rank=%d, alpha=%d)", config.lora.rank, config.lora.alpha)
     model = FastLanguageModel.get_peft_model(
         model,
-        r=lora_rank,
+        r=config.lora.rank,
-        lora_alpha=LORA_ALPHA,
+        lora_alpha=config.lora.alpha,
-        lora_dropout=LORA_DROPOUT,
+        lora_dropout=config.lora.dropout,
-        target_modules=LORA_TARGETS,
+        target_modules=config.lora.targets,
         bias="none",
         use_gradient_checkpointing="unsloth",
         random_state=42,
@@ -137,18 +156,18 @@ def main(
     logger.info("Split: %d train, %d validation", len(train_dataset), len(validation_dataset))
     training_args = TrainingArguments(
         output_dir=str(output_dir / "checkpoints"),
-        num_train_epochs=epochs,
+        num_train_epochs=config.training.epochs,
-        per_device_train_batch_size=batch_size,
+        per_device_train_batch_size=config.training.batch_size,
-        gradient_accumulation_steps=GRADIENT_ACCUMULATION,
+        gradient_accumulation_steps=config.training.gradient_accumulation,
-        learning_rate=learning_rate,
+        learning_rate=config.training.learning_rate,
-        warmup_ratio=WARMUP_RATIO,
+        warmup_ratio=config.training.warmup_ratio,
-        weight_decay=WEIGHT_DECAY,
+        weight_decay=config.training.weight_decay,
         lr_scheduler_type="cosine",
-        logging_steps=LOGGING_STEPS,
+        logging_steps=config.training.logging_steps,
-        save_steps=SAVE_STEPS,
+        save_steps=config.training.save_steps,
         save_total_limit=3,
         eval_strategy="steps",
-        eval_steps=SAVE_STEPS,
+        eval_steps=config.training.save_steps,
         load_best_model_at_end=True,
         bf16=True,
         optim="adamw_8bit",
@@ -162,11 +181,16 @@ def main(
         train_dataset=train_dataset,
         eval_dataset=validation_dataset,
         args=training_args,
-        max_seq_length=max_seq_length,
+        max_seq_length=config.training.max_seq_length,
         packing=True,
     )
 
-    logger.info("Starting training: %d train, %d val, %d epochs", len(train_dataset), len(validation_dataset), epochs)
+    logger.info(
+        "Starting training: %d train, %d val, %d epochs",
+        len(train_dataset),
+        len(validation_dataset),
+        config.training.epochs,
+    )
     trainer.train()
 
     merged_path = str(output_dir / "merged")