[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M75PD2VH645V9N4MQTCYGGEJ":3,"skill-same:01M75PD2VH645V9N4MQTCYGGEJ":99},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":51,"commitSha":52,"skillDirectory":53,"forksCount":14,"repositoryHtmlUrl":54,"repositoryHomepageUrl":6,"ownerHtmlUrl":55,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":56,"install":57,"resources":70},"01M75PD2VH645V9N4MQTCYGGEJ","fine-tuning-with-trl",null,"TRL: SFT, DPO, PPO, GRPO, reward modeling for LLM RLHF.","Iwakishirokoshu\u002FHermesUltimate","Iwakishirokoshu","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F194847626?v=4","01M75P4CVGA89TTCE8JGMNAQQ9","Around","01M75P4CWF85V6GVSEFKGGFZ03",0,"MIT","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: fine-tuning-with-trl\ndescription: \"TRL: SFT, DPO, PPO, GRPO, reward modeling for LLM RLHF.\"\nversion: 1.0.0\nauthor: Orchestra Research\nlicense: MIT\ndependencies: [trl, transformers, datasets, peft, accelerate, torch]\nplatforms: [linux, macos, windows]\nmetadata:\n  hermes:\n    tags: [Post-Training, TRL, Reinforcement Learning, Fine-Tuning, SFT, DPO, PPO, GRPO, RLHF, Preference Alignment, HuggingFace]\n\n---\n\n# TRL - Transformer Reinforcement Learning\n\n## Quick start\n\nTRL provides post-training methods for aligning language models with human preferences.\n\n**Installation**:\n```bash\npip install trl transformers datasets peft accelerate\n```\n\n**Supervised Fine-Tuning** (instruction tuning):\n```python\nfrom trl import SFTTrainer\n\ntrainer = SFTTrainer(\n    model=\"Qwen\u002FQwen2.5-0.5B\",\n    train_dataset=dataset,  # Prompt-completion pairs\n)\ntrainer.train()\n```\n\n**DPO** (align with preferences):\n```python\nfrom trl import DPOTrainer, DPOConfig\n\nconfig = DPOConfig(output_dir=\"model-dpo\", beta=0.1)\ntrainer = DPOTrainer(\n    model=model,\n    args=config,\n    train_dataset=preference_dataset,  # chosen\u002Frejected pairs\n    processing_class=tokenizer\n)\ntrainer.train()\n```\n\n## Common workflows\n\n### Workflow 1: Full RLHF pipeline (SFT → Reward Model → PPO)\n\nComplete pipeline from base model to human-aligned model.\n\nCopy this checklist:\n\n```\nRLHF Training:\n- [ ] Step 1: Supervised fine-tuning (SFT)\n- [ ] Step 2: Train reward model\n- [ ] Step 3: PPO reinforcement learning\n- [ ] Step 4: Evaluate aligned model\n```\n\n**Step 1: Supervised fine-tuning**\n\nTrain base model on instruction-following data:\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nfrom trl import SFTTrainer, SFTConfig\nfrom datasets import load_dataset\n\n# Load model\nmodel = AutoModelForCausalLM.from_pretrained(\"Qwen\u002FQwen2.5-0.5B\")\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen\u002FQwen2.5-0.5B\")\n\n# Load instruction dataset\ndataset = load_dataset(\"trl-lib\u002FCapybara\", split=\"train\")\n\n# Configure training\ntraining_args = SFTConfig(\n    output_dir=\"Qwen2.5-0.5B-SFT\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=2e-5,\n    logging_steps=10,\n    save_strategy=\"epoch\"\n)\n\n# Train\ntrainer = SFTTrainer(\n    model=model,\n    args=training_args,\n    train_dataset=dataset,\n    tokenizer=tokenizer\n)\ntrainer.train()\ntrainer.save_model()\n```\n\n**Step 2: Train reward model**\n\nTrain model to predict human preferences:\n\n```python\nfrom transformers import AutoModelForSequenceClassification\nfrom trl import RewardTrainer, RewardConfig\n\n# Load SFT model as base\nmodel = AutoModelForSequenceClassification.from_pretrained(\n    \"Qwen2.5-0.5B-SFT\",\n    num_labels=1  # Single reward score\n)\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen2.5-0.5B-SFT\")\n\n# Load preference data (chosen\u002Frejected pairs)\ndataset = load_dataset(\"trl-lib\u002Fultrafeedback_binarized\", split=\"train\")\n\n# Configure training\ntraining_args = RewardConfig(\n    output_dir=\"Qwen2.5-0.5B-Reward\",\n    per_device_train_batch_size=2,\n    num_train_epochs=1,\n    learning_rate=1e-5\n)\n\n# Train reward model\ntrainer = RewardTrainer(\n    model=model,\n    args=training_args,\n    processing_class=tokenizer,\n    train_dataset=dataset\n)\ntrainer.train()\ntrainer.save_model()\n```\n\n**Step 3: PPO reinforcement learning**\n\nOptimize policy using reward model:\n\n```bash\npython -m trl.scripts.ppo \\\n    --model_name_or_path Qwen2.5-0.5B-SFT \\\n    --reward_model_path Qwen2.5-0.5B-Reward \\\n    --dataset_name trl-internal-testing\u002Fdescriptiveness-sentiment-trl-style \\\n    --output_dir Qwen2.5-0.5B-PPO \\\n    --learning_rate 3e-6 \\\n    --per_device_train_batch_size 64 \\\n    --total_episodes 10000\n```\n\n**Step 4: Evaluate**\n\n```python\nfrom transformers import pipeline\n\n# Load aligned model\ngenerator = pipeline(\"text-generation\", model=\"Qwen2.5-0.5B-PPO\")\n\n# Test\nprompt = \"Explain quantum computing to a 10-year-old\"\noutput = generator(prompt, max_length=200)[0][\"generated_text\"]\nprint(output)\n```\n\n### Workflow 2: Simple preference alignment with DPO\n\nAlign model with preferences without reward model.\n\nCopy this checklist:\n\n```\nDPO Training:\n- [ ] Step 1: Prepare preference dataset\n- [ ] Step 2: Configure DPO\n- [ ] Step 3: Train with DPOTrainer\n- [ ] Step 4: Evaluate alignment\n```\n\n**Step 1: Prepare preference dataset**\n\nDataset format:\n```json\n{\n  \"prompt\": \"What is the capital of France?\",\n  \"chosen\": \"The capital of France is Paris.\",\n  \"rejected\": \"I don't know.\"\n}\n```\n\nLoad dataset:\n```python\nfrom datasets import load_dataset\n\ndataset = load_dataset(\"trl-lib\u002Fultrafeedback_binarized\", split=\"train\")\n# Or load your own\n# dataset = load_dataset(\"json\", data_files=\"preferences.json\")\n```\n\n**Step 2: Configure DPO**\n\n```python\nfrom trl import DPOConfig\n\nconfig = DPOConfig(\n    output_dir=\"Qwen2.5-0.5B-DPO\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=5e-7,\n    beta=0.1,  # KL penalty strength\n    max_prompt_length=512,\n    max_length=1024,\n    logging_steps=10\n)\n```\n\n**Step 3: Train with DPOTrainer**\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nfrom trl import DPOTrainer\n\nmodel = AutoModelForCausalLM.from_pretrained(\"Qwen\u002FQwen2.5-0.5B-Instruct\")\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen\u002FQwen2.5-0.5B-Instruct\")\n\ntrainer = DPOTrainer(\n    model=model,\n    args=config,\n    train_dataset=dataset,\n    processing_class=tokenizer\n)\n\ntrainer.train()\ntrainer.save_model()\n```\n\n**CLI alternative**:\n```bash\ntrl dpo \\\n    --model_name_or_path Qwen\u002FQwen2.5-0.5B-Instruct \\\n    --dataset_name argilla\u002FCapybara-Preferences \\\n    --output_dir Qwen2.5-0.5B-DPO \\\n    --per_device_train_batch_size 4 \\\n    --learning_rate 5e-7 \\\n    --beta 0.1\n```\n\n### Workflow 3: Memory-efficient online RL with GRPO\n\nTrain with reinforcement learning using minimal memory.\n\nFor in-depth GRPO guidance — reward function design, critical training insights (loss behavior, mode collapse, tuning), and advanced multi-stage patterns — see **[references\u002Fgrpo-training.md](references\u002Fgrpo-training.md)**. A production-ready training script is in **[templates\u002Fbasic_grpo_training.py](templates\u002Fbasic_grpo_training.py)**.\n\nCopy this checklist:\n\n```\nGRPO Training:\n- [ ] Step 1: Define reward function\n- [ ] Step 2: Configure GRPO\n- [ ] Step 3: Train with GRPOTrainer\n```\n\n**Step 1: Define reward function**\n\n```python\ndef reward_function(completions, **kwargs):\n    \"\"\"\n    Compute rewards for completions.\n\n    Args:\n        completions: List of generated texts\n\n    Returns:\n        List of reward scores (floats)\n    \"\"\"\n    rewards = []\n    for completion in completions:\n        # Example: reward based on length and unique words\n        score = len(completion.split())  # Favor longer responses\n        score += len(set(completion.lower().split()))  # Reward unique words\n        rewards.append(score)\n    return rewards\n```\n\nOr use a reward model:\n```python\nfrom transformers import pipeline\n\nreward_model = pipeline(\"text-classification\", model=\"reward-model-path\")\n\ndef reward_from_model(completions, prompts, **kwargs):\n    # Combine prompt + completion\n    full_texts = [p + c for p, c in zip(prompts, completions)]\n    # Get reward scores\n    results = reward_model(full_texts)\n    return [r[\"score\"] for r in results]\n```\n\n**Step 2: Configure GRPO**\n\n```python\nfrom trl import GRPOConfig\n\nconfig = GRPOConfig(\n    output_dir=\"Qwen2-GRPO\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=1e-5,\n    num_generations=4,  # Generate 4 completions per prompt\n    max_new_tokens=128\n)\n```\n\n**Step 3: Train with GRPOTrainer**\n\n```python\nfrom datasets import load_dataset\nfrom trl import GRPOTrainer\n\n# Load prompt-only dataset\ndataset = load_dataset(\"trl-lib\u002Ftldr\", split=\"train\")\n\ntrainer = GRPOTrainer(\n    model=\"Qwen\u002FQwen2-0.5B-Instruct\",\n    reward_funcs=reward_function,  # Your reward function\n    args=config,\n    train_dataset=dataset\n)\n\ntrainer.train()\n```\n\n**CLI**:\n```bash\ntrl grpo \\\n    --model_name_or_path Qwen\u002FQwen2-0.5B-Instruct \\\n    --dataset_name trl-lib\u002Ftldr \\\n    --output_dir Qwen2-GRPO \\\n    --num_generations 4\n```\n\n## When to use vs alternatives\n\n**Use TRL when:**\n- Need to align model with human preferences\n- Have preference data (chosen\u002Frejected pairs)\n- Want to use reinforcement learning (PPO, GRPO)\n- Need reward model training\n- Doing RLHF (full pipeline)\n\n**Method selection**:\n- **SFT**: Have prompt-completion pairs, want basic instruction following\n- **DPO**: Have preferences, want simple alignment (no reward model needed)\n- **PPO**: Have reward model, need maximum control over RL\n- **GRPO**: Memory-constrained, want online RL\n- **Reward Model**: Building RLHF pipeline, need to score generations\n\n**Use alternatives instead:**\n- **HuggingFace Trainer**: Basic fine-tuning without RL\n- **Axolotl**: YAML-based training configuration\n- **LitGPT**: Educational, minimal fine-tuning\n- **Unsloth**: Fast LoRA training\n\n## Common issues\n\n**Issue: OOM during DPO training**\n\nReduce batch size and sequence length:\n```python\nconfig = DPOConfig(\n    per_device_train_batch_size=1,  # Reduce from 4\n    max_length=512,  # Reduce from 1024\n    gradient_accumulation_steps=8  # Maintain effective batch\n)\n```\n\nOr use gradient checkpointing:\n```python\nmodel.gradient_checkpointing_enable()\n```\n\n**Issue: Poor alignment quality**\n\nTune beta parameter:\n```python\n# Higher beta = more conservative (stays closer to reference)\nconfig = DPOConfig(beta=0.5)  # Default 0.1\n\n# Lower beta = more aggressive alignment\nconfig = DPOConfig(beta=0.01)\n```\n\n**Issue: Reward model not learning**\n\nCheck loss type and learning rate:\n```python\nconfig = RewardConfig(\n    learning_rate=1e-5,  # Try different LR\n    num_train_epochs=3  # Train longer\n)\n```\n\nEnsure preference dataset has clear winners:\n```python\n# Verify dataset\nprint(dataset[0])\n# Should have clear chosen > rejected\n```\n\n**Issue: PPO training unstable**\n\nAdjust KL coefficient:\n```python\nconfig = PPOConfig(\n    kl_coef=0.1,  # Increase from 0.05\n    cliprange=0.1  # Reduce from 0.2\n)\n```\n\n## Advanced topics\n\n**SFT training guide**: See [references\u002Fsft-training.md](references\u002Fsft-training.md) for dataset formats, chat templates, packing strategies, and multi-GPU training.\n\n**DPO variants**: See [references\u002Fdpo-variants.md](references\u002Fdpo-variants.md) for IPO, cDPO, RPO, and other DPO loss functions with recommended hyperparameters.\n\n**Reward modeling**: See [references\u002Freward-modeling.md](references\u002Freward-modeling.md) for outcome vs process rewards, Bradley-Terry loss, and reward model evaluation.\n\n**Online RL methods**: See [references\u002Fonline-rl.md](references\u002Fonline-rl.md) for PPO, GRPO, RLOO, and OnlineDPO with detailed configurations.\n\n**GRPO deep dive**: See [references\u002Fgrpo-training.md](references\u002Fgrpo-training.md) for expert-level GRPO patterns — reward function design philosophy, training insights (why loss increases, mode collapse detection), hyperparameter tuning, multi-stage training, and troubleshooting. Production-ready template in [templates\u002Fbasic_grpo_training.py](templates\u002Fbasic_grpo_training.py).\n\n## Hardware requirements\n\n- **GPU**: NVIDIA (CUDA required)\n- **VRAM**: Depends on model and method\n  - SFT 7B: 16GB (with LoRA)\n  - DPO 7B: 24GB (stores reference model)\n  - PPO 7B: 40GB (policy + reward model)\n  - GRPO 7B: 24GB (more memory efficient)\n- **Multi-GPU**: Supported via `accelerate`\n- **Mixed precision**: BF16 recommended (A100\u002FH100)\n\n**Memory optimization**:\n- Use LoRA\u002FQLoRA for all methods\n- Enable gradient checkpointing\n- Use smaller batch sizes with gradient accumulation\n\n## Resources\n\n- Docs: https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Ftrl\u002F\n- GitHub: https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftrl\n- Papers:\n  - \"Training language models to follow instructions with human feedback\" (InstructGPT, 2022)\n  - \"Direct Preference Optimization: Your Language Model is Secretly a Reward Model\" (DPO, 2023)\n  - \"Group Relative Policy Optimization\" (GRPO, 2024)\n- Examples: https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftrl\u002Ftree\u002Fmain\u002Fexamples\u002Fscripts\n\n\n\n","# TRL - Transformer Reinforcement Learning\n\n## Quick start\n\nTRL provides post-training methods for aligning language models with human preferences.\n\n**Installation**:\n```bash\npip install trl transformers datasets peft accelerate\n```\n\n**Supervised Fine-Tuning** (instruction tuning):\n```python\nfrom trl import SFTTrainer\n\ntrainer = SFTTrainer(\n    model=\"Qwen\u002FQwen2.5-0.5B\",\n    train_dataset=dataset,  # Prompt-completion pairs\n)\ntrainer.train()\n```\n\n**DPO** (align with preferences):\n```python\nfrom trl import DPOTrainer, DPOConfig\n\nconfig = DPOConfig(output_dir=\"model-dpo\", beta=0.1)\ntrainer = DPOTrainer(\n    model=model,\n    args=config,\n    train_dataset=preference_dataset,  # chosen\u002Frejected pairs\n    processing_class=tokenizer\n)\ntrainer.train()\n```\n\n## Common workflows\n\n### Workflow 1: Full RLHF pipeline (SFT → Reward Model → PPO)\n\nComplete pipeline from base model to human-aligned model.\n\nCopy this checklist:\n\n```\nRLHF Training:\n- [ ] Step 1: Supervised fine-tuning (SFT)\n- [ ] Step 2: Train reward model\n- [ ] Step 3: PPO reinforcement learning\n- [ ] Step 4: Evaluate aligned model\n```\n\n**Step 1: Supervised fine-tuning**\n\nTrain base model on instruction-following data:\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nfrom trl import SFTTrainer, SFTConfig\nfrom datasets import load_dataset\n\n# Load model\nmodel = AutoModelForCausalLM.from_pretrained(\"Qwen\u002FQwen2.5-0.5B\")\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen\u002FQwen2.5-0.5B\")\n\n# Load instruction dataset\ndataset = load_dataset(\"trl-lib\u002FCapybara\", split=\"train\")\n\n# Configure training\ntraining_args = SFTConfig(\n    output_dir=\"Qwen2.5-0.5B-SFT\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=2e-5,\n    logging_steps=10,\n    save_strategy=\"epoch\"\n)\n\n# Train\ntrainer = SFTTrainer(\n    model=model,\n    args=training_args,\n    train_dataset=dataset,\n    tokenizer=tokenizer\n)\ntrainer.train()\ntrainer.save_model()\n```\n\n**Step 2: Train reward model**\n\nTrain model to predict human preferences:\n\n```python\nfrom transformers import AutoModelForSequenceClassification\nfrom trl import RewardTrainer, RewardConfig\n\n# Load SFT model as base\nmodel = AutoModelForSequenceClassification.from_pretrained(\n    \"Qwen2.5-0.5B-SFT\",\n    num_labels=1  # Single reward score\n)\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen2.5-0.5B-SFT\")\n\n# Load preference data (chosen\u002Frejected pairs)\ndataset = load_dataset(\"trl-lib\u002Fultrafeedback_binarized\", split=\"train\")\n\n# Configure training\ntraining_args = RewardConfig(\n    output_dir=\"Qwen2.5-0.5B-Reward\",\n    per_device_train_batch_size=2,\n    num_train_epochs=1,\n    learning_rate=1e-5\n)\n\n# Train reward model\ntrainer = RewardTrainer(\n    model=model,\n    args=training_args,\n    processing_class=tokenizer,\n    train_dataset=dataset\n)\ntrainer.train()\ntrainer.save_model()\n```\n\n**Step 3: PPO reinforcement learning**\n\nOptimize policy using reward model:\n\n```bash\npython -m trl.scripts.ppo \\\n    --model_name_or_path Qwen2.5-0.5B-SFT \\\n    --reward_model_path Qwen2.5-0.5B-Reward \\\n    --dataset_name trl-internal-testing\u002Fdescriptiveness-sentiment-trl-style \\\n    --output_dir Qwen2.5-0.5B-PPO \\\n    --learning_rate 3e-6 \\\n    --per_device_train_batch_size 64 \\\n    --total_episodes 10000\n```\n\n**Step 4: Evaluate**\n\n```python\nfrom transformers import pipeline\n\n# Load aligned model\ngenerator = pipeline(\"text-generation\", model=\"Qwen2.5-0.5B-PPO\")\n\n# Test\nprompt = \"Explain quantum computing to a 10-year-old\"\noutput = generator(prompt, max_length=200)[0][\"generated_text\"]\nprint(output)\n```\n\n### Workflow 2: Simple preference alignment with DPO\n\nAlign model with preferences without reward model.\n\nCopy this checklist:\n\n```\nDPO Training:\n- [ ] Step 1: Prepare preference dataset\n- [ ] Step 2: Configure DPO\n- [ ] Step 3: Train with DPOTrainer\n- [ ] Step 4: Evaluate alignment\n```\n\n**Step 1: Prepare preference dataset**\n\nDataset format:\n```json\n{\n  \"prompt\": \"What is the capital of France?\",\n  \"chosen\": \"The capital of France is Paris.\",\n  \"rejected\": \"I don't know.\"\n}\n```\n\nLoad dataset:\n```python\nfrom datasets import load_dataset\n\ndataset = load_dataset(\"trl-lib\u002Fultrafeedback_binarized\", split=\"train\")\n# Or load your own\n# dataset = load_dataset(\"json\", data_files=\"preferences.json\")\n```\n\n**Step 2: Configure DPO**\n\n```python\nfrom trl import DPOConfig\n\nconfig = DPOConfig(\n    output_dir=\"Qwen2.5-0.5B-DPO\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=5e-7,\n    beta=0.1,  # KL penalty strength\n    max_prompt_length=512,\n    max_length=1024,\n    logging_steps=10\n)\n```\n\n**Step 3: Train with DPOTrainer**\n\n```python\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\nfrom trl import DPOTrainer\n\nmodel = AutoModelForCausalLM.from_pretrained(\"Qwen\u002FQwen2.5-0.5B-Instruct\")\ntokenizer = AutoTokenizer.from_pretrained(\"Qwen\u002FQwen2.5-0.5B-Instruct\")\n\ntrainer = DPOTrainer(\n    model=model,\n    args=config,\n    train_dataset=dataset,\n    processing_class=tokenizer\n)\n\ntrainer.train()\ntrainer.save_model()\n```\n\n**CLI alternative**:\n```bash\ntrl dpo \\\n    --model_name_or_path Qwen\u002FQwen2.5-0.5B-Instruct \\\n    --dataset_name argilla\u002FCapybara-Preferences \\\n    --output_dir Qwen2.5-0.5B-DPO \\\n    --per_device_train_batch_size 4 \\\n    --learning_rate 5e-7 \\\n    --beta 0.1\n```\n\n### Workflow 3: Memory-efficient online RL with GRPO\n\nTrain with reinforcement learning using minimal memory.\n\nFor in-depth GRPO guidance — reward function design, critical training insights (loss behavior, mode collapse, tuning), and advanced multi-stage patterns — see **[references\u002Fgrpo-training.md](references\u002Fgrpo-training.md)**. A production-ready training script is in **[templates\u002Fbasic_grpo_training.py](templates\u002Fbasic_grpo_training.py)**.\n\nCopy this checklist:\n\n```\nGRPO Training:\n- [ ] Step 1: Define reward function\n- [ ] Step 2: Configure GRPO\n- [ ] Step 3: Train with GRPOTrainer\n```\n\n**Step 1: Define reward function**\n\n```python\ndef reward_function(completions, **kwargs):\n    \"\"\"\n    Compute rewards for completions.\n\n    Args:\n        completions: List of generated texts\n\n    Returns:\n        List of reward scores (floats)\n    \"\"\"\n    rewards = []\n    for completion in completions:\n        # Example: reward based on length and unique words\n        score = len(completion.split())  # Favor longer responses\n        score += len(set(completion.lower().split()))  # Reward unique words\n        rewards.append(score)\n    return rewards\n```\n\nOr use a reward model:\n```python\nfrom transformers import pipeline\n\nreward_model = pipeline(\"text-classification\", model=\"reward-model-path\")\n\ndef reward_from_model(completions, prompts, **kwargs):\n    # Combine prompt + completion\n    full_texts = [p + c for p, c in zip(prompts, completions)]\n    # Get reward scores\n    results = reward_model(full_texts)\n    return [r[\"score\"] for r in results]\n```\n\n**Step 2: Configure GRPO**\n\n```python\nfrom trl import GRPOConfig\n\nconfig = GRPOConfig(\n    output_dir=\"Qwen2-GRPO\",\n    per_device_train_batch_size=4,\n    num_train_epochs=1,\n    learning_rate=1e-5,\n    num_generations=4,  # Generate 4 completions per prompt\n    max_new_tokens=128\n)\n```\n\n**Step 3: Train with GRPOTrainer**\n\n```python\nfrom datasets import load_dataset\nfrom trl import GRPOTrainer\n\n# Load prompt-only dataset\ndataset = load_dataset(\"trl-lib\u002Ftldr\", split=\"train\")\n\ntrainer = GRPOTrainer(\n    model=\"Qwen\u002FQwen2-0.5B-Instruct\",\n    reward_funcs=reward_function,  # Your reward function\n    args=config,\n    train_dataset=dataset\n)\n\ntrainer.train()\n```\n\n**CLI**:\n```bash\ntrl grpo \\\n    --model_name_or_path Qwen\u002FQwen2-0.5B-Instruct \\\n    --dataset_name trl-lib\u002Ftldr \\\n    --output_dir Qwen2-GRPO \\\n    --num_generations 4\n```\n\n## When to use vs alternatives\n\n**Use TRL when:**\n- Need to align model with human preferences\n- Have preference data (chosen\u002Frejected pairs)\n- Want to use reinforcement learning (PPO, GRPO)\n- Need reward model training\n- Doing RLHF (full pipeline)\n\n**Method selection**:\n- **SFT**: Have prompt-completion pairs, want basic instruction following\n- **DPO**: Have preferences, want simple alignment (no reward model needed)\n- **PPO**: Have reward model, need maximum control over RL\n- **GRPO**: Memory-constrained, want online RL\n- **Reward Model**: Building RLHF pipeline, need to score generations\n\n**Use alternatives instead:**\n- **HuggingFace Trainer**: Basic fine-tuning without RL\n- **Axolotl**: YAML-based training configuration\n- **LitGPT**: Educational, minimal fine-tuning\n- **Unsloth**: Fast LoRA training\n\n## Common issues\n\n**Issue: OOM during DPO training**\n\nReduce batch size and sequence length:\n```python\nconfig = DPOConfig(\n    per_device_train_batch_size=1,  # Reduce from 4\n    max_length=512,  # Reduce from 1024\n    gradient_accumulation_steps=8  # Maintain effective batch\n)\n```\n\nOr use gradient checkpointing:\n```python\nmodel.gradient_checkpointing_enable()\n```\n\n**Issue: Poor alignment quality**\n\nTune beta parameter:\n```python\n# Higher beta = more conservative (stays closer to reference)\nconfig = DPOConfig(beta=0.5)  # Default 0.1\n\n# Lower beta = more aggressive alignment\nconfig = DPOConfig(beta=0.01)\n```\n\n**Issue: Reward model not learning**\n\nCheck loss type and learning rate:\n```python\nconfig = RewardConfig(\n    learning_rate=1e-5,  # Try different LR\n    num_train_epochs=3  # Train longer\n)\n```\n\nEnsure preference dataset has clear winners:\n```python\n# Verify dataset\nprint(dataset[0])\n# Should have clear chosen > rejected\n```\n\n**Issue: PPO training unstable**\n\nAdjust KL coefficient:\n```python\nconfig = PPOConfig(\n    kl_coef=0.1,  # Increase from 0.05\n    cliprange=0.1  # Reduce from 0.2\n)\n```\n\n## Advanced topics\n\n**SFT training guide**: See [references\u002Fsft-training.md](references\u002Fsft-training.md) for dataset formats, chat templates, packing strategies, and multi-GPU training.\n\n**DPO variants**: See [references\u002Fdpo-variants.md](references\u002Fdpo-variants.md) for IPO, cDPO, RPO, and other DPO loss functions with recommended hyperparameters.\n\n**Reward modeling**: See [references\u002Freward-modeling.md](references\u002Freward-modeling.md) for outcome vs process rewards, Bradley-Terry loss, and reward model evaluation.\n\n**Online RL methods**: See [references\u002Fonline-rl.md](references\u002Fonline-rl.md) for PPO, GRPO, RLOO, and OnlineDPO with detailed configurations.\n\n**GRPO deep dive**: See [references\u002Fgrpo-training.md](references\u002Fgrpo-training.md) for expert-level GRPO patterns — reward function design philosophy, training insights (why loss increases, mode collapse detection), hyperparameter tuning, multi-stage training, and troubleshooting. Production-ready template in [templates\u002Fbasic_grpo_training.py](templates\u002Fbasic_grpo_training.py).\n\n## Hardware requirements\n\n- **GPU**: NVIDIA (CUDA required)\n- **VRAM**: Depends on model and method\n  - SFT 7B: 16GB (with LoRA)\n  - DPO 7B: 24GB (stores reference model)\n  - PPO 7B: 40GB (policy + reward model)\n  - GRPO 7B: 24GB (more memory efficient)\n- **Multi-GPU**: Supported via `accelerate`\n- **Mixed precision**: BF16 recommended (A100\u002FH100)\n\n**Memory optimization**:\n- Use LoRA\u002FQLoRA for all methods\n- Enable gradient checkpointing\n- Use smaller batch sizes with gradient accumulation\n\n## Resources\n\n- Docs: https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Ftrl\u002F\n- GitHub: https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftrl\n- Papers:\n  - \"Training language models to follow instructions with human feedback\" (InstructGPT, 2022)\n  - \"Direct Preference Optimization: Your Language Model is Secretly a Reward Model\" (DPO, 2023)\n  - \"Group Relative Policy Optimization\" (GRPO, 2024)\n- Examples: https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftrl\u002Ftree\u002Fmain\u002Fexamples\u002Fscripts\n\n\n\n",{"name":5,"author":24,"license":15,"version":25,"metadata":26,"platforms":40,"description":7,"dependencies":44},"Orchestra Research","1.0.0",{"hermes":27},{"tags":28},[29,30,31,32,33,34,35,36,37,38,39],"Post-Training","TRL","Reinforcement Learning","Fine-Tuning","SFT","DPO","PPO","GRPO","RLHF","Preference Alignment","HuggingFace",[41,42,43],"linux","macos","windows",[45,46,47,48,49,50],"trl","transformers","datasets","peft","accelerate","torch","main","c0c483827a963764d5e1a663a9591867d47289e6","optional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu","2026-06-15 13:48:41",{"source":8,"skillName":5,"commands":58,"prompt":68,"zipAvailable":69},[59,62,65],{"tool":60,"command":61},"npx","npx skills add Iwakishirokoshu\u002FHermesUltimate --skill fine-tuning-with-trl",{"tool":63,"command":64},"bunx","bunx skills add Iwakishirokoshu\u002FHermesUltimate --skill fine-tuning-with-trl",{"tool":66,"command":67},"pnpm","pnpm dlx skills add Iwakishirokoshu\u002FHermesUltimate --skill fine-tuning-with-trl","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002FSKILL.md",true,[71,77,81,85,89,93],{"path":72,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":75,"sourceHtmlUrl":76},"references\u002Fdpo-variants.md","REFERENCE","参考文档",4297,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Freferences\u002Fdpo-variants.md",{"path":78,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":79,"sourceHtmlUrl":80},"references\u002Fgrpo-training.md",15901,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Freferences\u002Fgrpo-training.md",{"path":82,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":83,"sourceHtmlUrl":84},"references\u002Fonline-rl.md",1971,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Freferences\u002Fonline-rl.md",{"path":86,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Freward-modeling.md",2597,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Freferences\u002Freward-modeling.md",{"path":90,"resourceType":73,"resourceTypeDesc":74,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Fsft-training.md",3237,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Freferences\u002Fsft-training.md",{"path":94,"resourceType":95,"resourceTypeDesc":96,"sizeBytes":97,"sourceHtmlUrl":98},"templates\u002Fbasic_grpo_training.py","SCRIPT","脚本",6113,"https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Foptional-skills\u002Fmlops\u002Ftraining\u002Ftrl-fine-tuning\u002Ftemplates\u002Fbasic_grpo_training.py",[100,107,114,121,128,135],{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":19},"01M770W8JDE5EC1TJT4GSDWD78","wps-pixie-dust","WPS Pixie-Dust offline nonce attack (reaver -K \u002F pixiewps) and fallback online PIN brute (bully) to recover the AP's WPA PSK without capturing a handshake.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwps-pixie-dust\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":113,"publishedAt":19},"01M770W4B7T5VCVHB88V1F02H1","wpa3-sae","WPA3-SAE transition-mode downgrade (DragonShift), SSID Confusion CVE-2023-52424, Dragonblood side-channels, and SAE captive-portal credential recovery against WPA3-Personal networks.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa3-sae\u002FSKILL.md",[],[],{"skillNo":115,"name":116,"nameZh":6,"description":117,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":118,"categories":119,"keywords":120,"publishedAt":19},"01M770W093C8FAV8S34K8VB6T9","wpa2-psk","WPA\u002FWPA2-PSK handshake capture via targeted deauth + PMKID (no deauth required) + offline hashcat cracking. The most common consumer\u002FSMB encryption mode in 2026.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa2-psk\u002FSKILL.md",[],[],{"skillNo":122,"name":123,"nameZh":6,"description":124,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":125,"categories":126,"keywords":127,"publishedAt":19},"01M770VTNAQGZ3D0RQVFEKTE8Z","wpa-enterprise-eap","WPA\u002FWPA2\u002FWPA3-Enterprise (802.1X\u002FEAP) rogue-RADIUS evil-twin for MSCHAPv2 capture, GTC downgrade, and PEAP relay. MSCHAPv2 capture equals a NetNTLM hash — the primary wireless on-ramp to Active Directory.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fwpa-enterprise-eap\u002FSKILL.md",[],[],{"skillNo":129,"name":130,"nameZh":6,"description":131,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":132,"categories":133,"keywords":134,"publishedAt":19},"01M770VNG05ZX3CV84R9GX8PJN","krack-fragattacks","KRACK key-reinstallation (CVE-2017-13077..13082) and FragAttacks fragmentation\u002Faggregation flaws (CVE-2020-24586..24588, CVE-2020-26139..26147) against legacy or embedded 802.11 supplicants with poor patch cadence.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fkrack-fragattacks\u002FSKILL.md",[],[],{"skillNo":136,"name":137,"nameZh":6,"description":138,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":139,"categories":140,"keywords":141,"publishedAt":19},"01M770VHFH7N192V56PFSRXF5K","evil-twin-karma","Evil-twin rogue AP with KARMA\u002FMana PNL-probe response, captive-portal credential capture, and post-association MITM for PSK\u002Fopen networks. Distinct from wpa-enterprise-eap which targets 802.1X.","https:\u002F\u002Fgithub.com\u002FIwakishirokoshu\u002FHermesUltimate\u002Fblob\u002Fmain\u002Fstack\u002Fdecepticon-slim\u002Fupstream\u002Fpackages\u002Fdecepticon\u002Fdecepticon\u002Fskills\u002Fstandard\u002Fwireless\u002Fevil-twin-karma\u002FSKILL.md",[],[]]