Pular para o conteúdo principal

​
Templates de Configuração

Copie esses templates e personalize para seu caso de uso.

​
Templates de Treinamento LLM

​
SFT Rápido (Modelo Pequeno)

Melhor para testes e datasets pequenos: 
task: llm-sft
backend: local
base_model: google/gemma-3-270m
project_name: quick-sft

data:
  path: ./data.jsonl
  train_split: train
  valid_split: null
  chat_template: tokenizer
  column_mapping:
    text_column: text

log: wandb

params:
  epochs: 3
  batch_size: 4
  lr: 3e-5

​
SFT de Produção (Modelo 7B)

Para fine-tuning completo com LoRA: 
task: llm-sft
backend: local
base_model: meta-llama/Llama-3.2-8B
project_name: production-sft

data:
  path: ./conversations.jsonl
  train_split: train
  valid_split: validation
  chat_template: tokenizer
  column_mapping:
    text_column: text

log: wandb

hub:
  push_to_hub: false

params:
  epochs: 3
  batch_size: 2
  gradient_accumulation: 8
  lr: 3e-5
  warmup_ratio: 0.1
  mixed_precision: bf16

  peft: true
  lora_r: 32
  lora_alpha: 64
  lora_dropout: 0.05

  use_flash_attention_2: true
  packing: true

  save_strategy: steps
  save_steps: 500
  save_total_limit: 1

​
Treinamento DPO de Preferências

Para alinhamento de preferências: 
task: llm-dpo
backend: local
base_model: meta-llama/Llama-3.2-1B
project_name: dpo-aligned

data:
  path: ./preferences.jsonl
  train_split: train
  valid_split: null
  chat_template: tokenizer
  column_mapping:
    prompt_text_column: prompt
    text_column: chosen
    rejected_text_column: rejected

log: wandb

params:
  dpo_beta: 0.1
  max_prompt_length: 128
  max_completion_length: null

  epochs: 1
  batch_size: 2
  gradient_accumulation: 4
  lr: 5e-6

  peft: true
  lora_r: 16
  lora_alpha: 32

​
Treinamento ORPO

Otimização combinada SFT + preferências: 
task: llm-orpo
backend: local
base_model: google/gemma-2-2b
project_name: gemma-orpo

data:
  path: ./preferences.jsonl
  train_split: train
  valid_split: null
  chat_template: tokenizer
  column_mapping:
    prompt_text_column: prompt
    text_column: chosen
    rejected_text_column: rejected

log: wandb

params:
  dpo_beta: 0.1
  max_prompt_length: 128
  max_completion_length: null

  epochs: 3
  batch_size: 2
  lr: 5e-5

  peft: true
  lora_r: 16

​
Distilação de Conhecimento

Para comprimir modelos maiores: 
task: llm-sft
backend: local
base_model: google/gemma-3-270m
project_name: distilled-gemma

data:
  path: ./prompts.jsonl
  train_split: train
  valid_split: null
  chat_template: tokenizer
  column_mapping:
    text_column: text

log: wandb

params:
  use_distillation: true
  teacher_model: google/gemma-2-2b
  distill_temperature: 3.0
  distill_alpha: 0.7

  epochs: 5
  batch_size: 8
  lr: 1e-4

​
Templates de Classificação

​
Classificação de Texto

task: text-classification
backend: local
base_model: bert-base-uncased
project_name: sentiment

data:
  path: ./reviews.csv
  train_split: train
  valid_split: null
  column_mapping:
    text_column: text
    target_column: target

log: wandb

params:
  epochs: 5
  batch_size: 16
  lr: 5e-5

​
Classificação Multi-Classe

task: text-classification
backend: local
base_model: microsoft/deberta-v3-base
project_name: categorizer

data:
  path: ./categories.csv
  train_split: train
  valid_split: validation
  column_mapping:
    text_column: content
    target_column: target

log: wandb

params:
  epochs: 10
  batch_size: 8
  lr: 1e-5
  warmup_ratio: 0.1

​
Classificação de Tokens (NER)

task: token-classification
backend: local
base_model: bert-base-cased
project_name: entity-extractor

data:
  path: ./ner_data.json
  train_split: train
  valid_split: null
  column_mapping:
    tokens_column: tokens
    tags_column: tags

log: wandb

params:
  epochs: 5
  batch_size: 16
  lr: 5e-5

​
Templates de Visão

​
Classificação de Imagem

task: image-classification
backend: local
base_model: google/vit-base-patch16-224
project_name: image-classifier

data:
  path: ./images/
  train_split: train
  valid_split: null
  column_mapping:
    image_column: image
    target_column: target

log: wandb

params:
  epochs: 10
  batch_size: 32
  lr: 5e-5

​
Detecção de Objetos

task: object-detection
backend: local
base_model: facebook/detr-resnet-50
project_name: detector

data:
  path: ./coco_format/
  train_split: train
  valid_split: null
  column_mapping:
    image_column: image
    objects_column: objects

log: wandb

params:
  epochs: 20
  batch_size: 8
  lr: 1e-4

​
Modelo Visão-Linguagem (VQA)

task: vlm:vqa
backend: local
base_model: google/paligemma-3b-pt-224
project_name: vlm-vqa

data:
  path: ./vqa_data.jsonl
  train_split: train
  valid_split: null
  column_mapping:
    image_column: image
    text_column: text
    prompt_text_column: prompt

log: wandb

params:
  epochs: 3
  batch_size: 2
  lr: 5e-5
  gradient_accumulation: 4

  peft: true
  lora_r: 16
  lora_alpha: 32

​
Templates Avançados

​
Varredura de Hiperparâmetros

task: llm-sft
backend: local
base_model: google/gemma-3-270m
project_name: sweep-experiment

data:
  path: ./data.jsonl
  train_split: train
  valid_split: validation
  chat_template: tokenizer
  column_mapping:
    text_column: text

log: wandb

params:
  use_sweep: true
  sweep_backend: optuna
  sweep_n_trials: 20
  sweep_metric: eval_loss
  sweep_direction: minimize

  # Base parameters (sweep will vary these)
  epochs: 3
  batch_size: 4
  lr: 3e-5

  peft: true
  lora_r: 16

​
Sentence Transformers

task: sentence-transformers:pair_score
backend: local
base_model: sentence-transformers/all-MiniLM-L6-v2
project_name: embeddings

data:
  path: ./pairs.csv
  train_split: train
  valid_split: null
  column_mapping:
    sentence1_column: sentence1
    sentence2_column: sentence2
    target_column: score

log: wandb

params:
  epochs: 3
  batch_size: 8
  lr: 3e-5

​
Uso

Salve qualquer template como config.yaml e execute: 
aitraining --config config.yaml

​
Próximos Passos

Configurações YAML

Detalhes da estrutura do arquivo de configuração

Treinamento LLM

Referência de parâmetros