SmolFactory / config /train_smollm3_openhermes_fr.py
Tonic's picture
adds formatting fix
ebe598e verified
raw
history blame
4.48 kB
"""
SmolLM3 Training Configuration for OpenHermes-FR Dataset
Optimized for French instruction tuning using legmlai/openhermes-fr
"""
import os
from dataclasses import dataclass
from typing import Optional
from config.train_smollm3 import SmolLM3Config
@dataclass
class SmolLM3ConfigOpenHermesFR(SmolLM3Config):
"""Configuration for SmolLM3 fine-tuning on OpenHermes-FR dataset"""
# Model configuration
model_name: str = "HuggingFaceTB/SmolLM3-3B"
max_seq_length: int = 4096
use_flash_attention: bool = True
use_gradient_checkpointing: bool = True
# Training configuration - optimized for French instruction tuning
batch_size: int = 2 # Reduced for French text (longer sequences)
gradient_accumulation_steps: int = 8 # Increased to maintain effective batch size
learning_rate: float = 1e-5 # Slightly lower for instruction tuning
weight_decay: float = 0.01
warmup_steps: int = 500 # More warmup for instruction tuning
max_iters: int = 2000 # More iterations for large dataset
eval_interval: int = 200
log_interval: int = 10
save_interval: int = 500
# Optimizer configuration
optimizer: str = "adamw_torch"
beta1: float = 0.9
beta2: float = 0.95
eps: float = 1e-8
# Scheduler configuration
scheduler: str = "cosine"
min_lr: float = 1e-6
# Mixed precision
fp16: bool = True
bf16: bool = False
# DDP configuration
ddp_backend: str = "nccl"
ddp_find_unused_parameters: bool = False
# Logging and saving
save_steps: int = 500
eval_steps: int = 200
logging_steps: int = 10
save_total_limit: Optional[int] = 3
# Evaluation
eval_strategy: str = "steps"
metric_for_best_model: str = "eval_loss"
greater_is_better: bool = False
load_best_model_at_end: bool = True
# OpenHermes-FR Dataset configuration
dataset_name: str = "legmlai/openhermes-fr"
dataset_split: str = "train"
input_field: str = "prompt"
target_field: str = "accepted_completion"
filter_bad_entries: bool = True
bad_entry_field: str = "bad_entry"
# Data configuration (not used for HF datasets but kept for compatibility)
data_dir: str = None
train_file: str = None
validation_file: Optional[str] = None
test_file: Optional[str] = None
# Chat template configuration
use_chat_template: bool = True
chat_template_kwargs: dict = None
# Trackio monitoring configuration
enable_tracking: bool = True
trackio_url: Optional[str] = None
trackio_token: Optional[str] = None
log_artifacts: bool = True
log_metrics: bool = True
log_config: bool = True
experiment_name: Optional[str] = None
# HF Datasets configuration
hf_token: Optional[str] = None
dataset_repo: Optional[str] = None
def __post_init__(self):
if self.chat_template_kwargs is None:
self.chat_template_kwargs = {
"add_generation_prompt": True,
"no_think_system_message": True # Set to True to add /no_think tag
}
# Validate configuration
if self.fp16 and self.bf16:
raise ValueError("Cannot use both fp16 and bf16")
if self.max_seq_length > 131072: # 128k limit
raise ValueError("max_seq_length cannot exceed 131072")
# Set default experiment name if not provided
if self.experiment_name is None:
self.experiment_name = "smollm3_openhermes_fr"
def get_config(config_path: str) -> SmolLM3ConfigOpenHermesFR:
"""Load configuration from file or return default"""
if os.path.exists(config_path):
# Load from file if it exists
import importlib.util
spec = importlib.util.spec_from_file_location("config_module", config_path)
config_module = importlib.util.module_from_spec(spec)
spec.loader.exec_module(config_module)
if hasattr(config_module, 'config'):
return config_module.config
else:
# Try to find a config class
for attr_name in dir(config_module):
attr = getattr(config_module, attr_name)
if isinstance(attr, SmolLM3ConfigOpenHermesFR):
return attr
# Return default configuration
return SmolLM3ConfigOpenHermesFR()
# Default configuration instance
config = SmolLM3ConfigOpenHermesFR()