Hugging Face Model Trainer

TRL Training on Hugging Face Jobs

Overview

Train language models using TRL (Transformer Reinforcement Learning) on fully managed Hugging Face infrastructure. No local GPU setup required—models train on cloud GPUs and results are automatically saved to the Hugging Face Hub.

TRL provides multiple training methods:

• SFT (Supervised Fine-Tuning) - Standard instruction tuning
• DPO (Direct Preference Optimization) - Alignment from preference data
• GRPO (Group Relative Policy Optimization) - Online RL training
• Reward Modeling - Train reward models for RLHF

For detailed TRL method documentation:

hf_doc_search("your query", product="trl")
hf_doc_fetch("https://huggingface.co/docs/trl/sft_trainer")  # SFT
hf_doc_fetch("https://huggingface.co/docs/trl/dpo_trainer")  # DPO
# etc.

See also: references/training_methods.md for method overviews and selection guidance

When to Use This Skill

Use this skill when users want to:

• Fine-tune language models on cloud GPUs without local infrastructure
• Train with TRL methods (SFT, DPO, GRPO, etc.)
• Run training jobs on Hugging Face Jobs infrastructure
• Convert trained models to GGUF for local deployment (Ollama, LM Studio, llama.cpp)
• Ensure trained models are permanently saved to the Hub
• Use modern workflows with optimized defaults

When to Use Unsloth

Use Unsloth (references/unsloth.md) instead of standard TRL when:

• Limited GPU memory - Unsloth uses ~60% less VRAM
• Speed matters - Unsloth is ~2x faster
• Training large models (>13B) - memory efficiency is critical
• Training Vision-Language Models (VLMs) - Unsloth has FastVisionModel support

See references/unsloth.md for complete Unsloth documentation and scripts/unsloth_sft_example.py for a production-ready training script.

Key Directives

When assisting with training jobs:

1. ALWAYS use hf_jobs() MCP tool - Submit jobs using hf_jobs("uv", {...}), NOT bash trl-jobs commands. The script parameter accepts Python code directly. Do NOT save to local files unless the user explicitly requests it. Pass the script content as a string to hf_jobs(). If user asks to "train a model", "fine-tune", or similar requests, you MUST create the training script AND submit the job immediately using hf_jobs().

1. Always include Trackio - Every training script should include Trackio for real-time monitoring. Use example scripts in scripts/ as templates.

1. Provide job details after submission - After submitting, provide job ID, monitoring URL, estimated time, and note that the user can request status checks later.

1. Use example scripts as templates - Reference scripts/train_sft_example.py, scripts/train_dpo_example.py, etc. as starting points.

Local Script Dependencies

To run scripts locally (like estimate_cost.py), install dependencies:

pip install -r requirements.txt

Prerequisites Checklist

Before starting any training job, verify:

✅ Account & Authentication

• Hugging Face Account with Pro, Team, or Enterprise plan (Jobs require paid plan)
• Authenticated login: Check with hf_whoami()
• HF_TOKEN for Hub Push ⚠️ CRITICAL - Training environment is ephemeral, must push to Hub or ALL training results are lost
• Token must have write permissions
• MUST pass secrets={"HF_TOKEN": "$HF_TOKEN"} in job config to make token available (the $HF_TOKEN syntax

references your actual token value)

✅ Dataset Requirements

• Dataset must exist on Hub or be loadable via datasets.load_dataset()
• Format must match training method (SFT: "messages"/text/prompt-completion; DPO: chosen/rejected; GRPO: prompt-only)
• ALWAYS validate unknown datasets before GPU training to prevent format failures (see Dataset Validation section below)
• Size appropriate for hardware (Demo: 50-100 examples on t4-small; Production: 1K-10K+ on a10g-large/a100-large)

⚠️ Critical Settings

• Timeout must exceed expected training time - Default 30min is TOO SHORT for most training. Minimum recommended: 1-2 hours. Job fails and loses all progress if timeout is exceeded.
• Hub push must be enabled - Config: push_to_hub=True, hub_model_id="username/model-name"; Job: secrets={"HF_TOKEN": "$HF_TOKEN"}

Asynchronous Job Guidelines

⚠️ IMPORTANT: Training jobs run asynchronously and can take hours

Action Required

When user requests training:

1. Create the training script with Trackio included (use scripts/train_sft_example.py as template)
2. Submit immediately using hf_jobs() MCP tool with script content inline - don't save to file unless user requests
3. Report submission with job ID, monitoring URL, and estimated time
4. Wait for user to request status checks - don't poll automatically

Ground Rules

• Jobs run in background - Submission returns immediately; training continues independently
• Initial logs delayed - Can take 30-60 seconds for logs to appear
• User checks status - Wait for user to request status updates
• Avoid polling - Check logs only on user request; provide monitoring links instead

After Submission

Provide to user:

• ✅ Job ID and monitoring URL
• ✅ Expected completion time
• ✅ Trackio dashboard URL
• ✅ Note that user can request status checks later

Example Response:

✅ Job submitted successfully!

Job ID: abc123xyz
Monitor: https://huggingface.co/jobs/username/abc123xyz

Expected time: ~2 hours
Estimated cost: ~$10

The job is running in the background. Ask me to check status/logs when ready!

Quick Start: Three Approaches

💡 Tip for Demos: For quick demos on smaller GPUs (t4-small), omit eval_dataset and eval_strategy to save ~40% memory. You'll still see training loss and learning progress.

Sequence Length Configuration

TRL config classes use max_length (not max_seq_length) to control tokenized sequence length:

# ✅ CORRECT - If you need to set sequence length
SFTConfig(max_length=512)   # Truncate sequences to 512 tokens
DPOConfig(max_length=2048)  # Longer context (2048 tokens)

# ❌ WRONG - This parameter doesn't exist
SFTConfig(max_seq_length=512)  # TypeError!

Default behavior: max_length=1024 (truncates from right). This works well for most training.

When to override:

• Longer context: Set higher (e.g., max_length=2048)
• Memory constraints: Set lower (e.g., max_length=512)
• Vision models: Set max_length=None (prevents cutting image tokens)

Usually you don't need to set this parameter at all - the examples below use the sensible default.

Approach 1: UV Scripts (Recommended—Default Choice)

UV scripts use PEP 723 inline dependencies for clean, self-contained training. This is the primary approach for Claude Code.

hf_jobs("uv", {
    "script": """
# /// script
# dependencies = ["trl>=0.12.0", "peft>=0.7.0", "trackio"]
# ///

from datasets import load_dataset
from peft import LoraConfig
from trl import SFTTrainer, SFTConfig
import trackio

dataset = load_dataset("trl-lib/Capybara", split="train")

# Create train/eval split for monitoring
dataset_split = dataset.train_test_split(test_size=0.1, seed=42)

trainer = SFTTrainer(
    model="Qwen/Qwen2.5-0.5B",
    train_dataset=dataset_split["train"],
    eval_dataset=dataset_split["test"],
    peft_config=LoraConfig(r=16, lora_alpha=32),
    args=SFTConfig(
        output_dir="my-model",
        push_to_hub=True,
        hub_model_id="username/my-model",
        num_train_epochs=3,
        eval_strategy="steps",
        eval_steps=50,
        report_to="trackio",
        project="meaningful_prject_name", # project name for the training name (trackio)
        run_name="meaningful_run_name",   # descriptive name for the specific training run (trackio)
    )
)

trainer.train()
trainer.push_to_hub()
""",
    "flavor": "a10g-large",
    "timeout": "2h",
    "secrets": {"HF_TOKEN": "$HF_TOKEN"}
})

Benefits: Direct MCP tool usage, clean code, dependencies declared inline (PEP 723), no file saving required, full control When to use: Default choice for all training tasks in Claude Code, custom training logic, any scenario requiring hf_jobs()

Working with Scripts

⚠️ Important: The script parameter accepts either inline code (as shown above) OR a URL. Local file paths do NOT work.

Why local paths don't work: Jobs run in isolated Docker containers without access to your local filesystem. Scripts must be:

• Inline code (recommended for custom training)
• Publicly accessible URLs
• Private repo URLs (with HF_TOKEN)

Common mistakes:

# ❌ These will all fail
hf_jobs("uv", {"script": "train.py"})
hf_jobs("uv", {"script": "./scripts/train.py"})
hf_jobs("uv", {"script": "/path/to/train.py"})

Correct approaches:

# ✅ Inline code (recommended)
hf_jobs("uv", {"script": "# /// script\n# dependencies = [...]\n# ///\n\n"})

# ✅ From Hugging Face Hub
hf_jobs("uv", {"script": "https://huggingface.co/user/repo/resolve/main/train.py"})

# ✅ From GitHub
hf_jobs("uv", {"script": "https://raw.githubusercontent.com/user/repo/main/train.py"})

# ✅ From Gist
hf_jobs("uv", {"script": "https://gist.githubusercontent.com/user/id/raw/train.py"})

To use local scripts: Upload to HF Hub first:

huggingface-cli repo create my-training-scripts --type model
huggingface-cli upload my-training-scripts ./train.py train.py
# Use: https://huggingface.co/USERNAME/my-training-scripts/resolve/main/train.py

Approach 2: TRL Maintained Scripts (Official Examples)

TRL provides battle-tested scripts for all methods. Can be run from URLs:

hf_jobs("uv", {
    "script": "https://github.com/huggingface/trl/blob/main/trl/scripts/sft.py",
    "script_args": [
        "--model_name_or_path", "Qwen/Qwen2.5-0.5B",
        "--dataset_name", "trl-lib/Capybara",
        "--output_dir", "my-model",
        "--push_to_hub",
        "--hub_model_id", "username/my-model"
    ],
    "flavor": "a10g-large",
    "timeout": "2h",
    "secrets": {"HF_TOKEN": "$HF_TOKEN"}
})

Benefits: No code to write, maintained by TRL team, production-tested When to use: Standard TRL training, quick experiments, don't need custom code Available: Scripts are available from https://github.com/huggingface/trl/tree/main/examples/scripts

Finding More UV Scripts on Hub

The uv-scripts organization provides ready-to-use UV scripts stored as datasets on Hugging Face Hub:

# Discover available UV script collections
dataset_search({"author": "uv-scripts", "sort": "downloads", "limit": 20})

# Explore a specific collection
hub_repo_details(["uv-scripts/classification"], repo_type="dataset", include_readme=True)

Popular collections: ocr, classification, synthetic-data, vllm, dataset-creation

Approach 3: HF Jobs CLI (Direct Terminal Commands)

When the hf_jobs() MCP tool is unavailable, use the hf jobs CLI directly.

⚠️ CRITICAL: CLI Syntax Rules

# ✅ CORRECT syntax - flags BEFORE script URL
hf jobs uv run --flavor a10g-large --timeout 2h --secrets HF_TOKEN "https://example.com/train.py"

# ❌ WRONG - "run uv" instead of "uv run"
hf jobs run uv "https://example.com/train.py" --flavor a10g-large

# ❌ WRONG - flags AFTER script URL (will be ignored!)
hf jobs uv run "https://example.com/train.py" --flavor a10g-large

# ❌ WRONG - "--secret" instead of "--secrets" (plural)
hf jobs uv run --secret HF_TOKEN "https://example.com/train.py"

Key syntax rules:

1. Command order is hf jobs uv run (NOT hf jobs run uv)
2. All flags (--flavor, --timeout, --secrets) must come BEFORE the script URL
3. Use --secrets (plural), not --secret
4. Script URL must be the last positional argument

Complete CLI example:

hf jobs uv run \
  --flavor a10g-large \
  --timeout 2h \
  --secrets HF_TOKEN \
  "https://huggingface.co/user/repo/resolve/main/train.py"

Check job status via CLI:

hf jobs ps                        # List all jobs
hf jobs logs              # View logs
hf jobs inspect           # Job details
hf jobs cancel            # Cancel a job

Approach 4: TRL Jobs Package (Simplified Training)

The trl-jobs package provides optimized defaults and one-liner training.

# Install
pip install trl-jobs

# Train with SFT (simplest possible)
trl-jobs sft \
  --model_name Qwen/Qwen2.5-0.5B \
  --dataset_name trl-lib/Capybara

Benefits: Pre-configured settings, automatic Trackio integration, automatic Hub push, one-line commands When to use: User working in terminal directly (not Claude Code context), quick local experimentation Repository: https://github.com/huggingface/trl-jobs

⚠️ In Claude Code context, prefer using hf_jobs() MCP tool (Approach 1) when available.

Hardware Selection

| Model Size | Recommended Hardware | Cost (approx/hr) | Use Case | |------------|---------------------|------------------|----------| | 7B to reduce memory

• Multi-GPU automatically handled by TRL/Accelerate
• Start with smaller hardware for testing

See: references/hardware_guide.md for detailed specifications

Critical: Saving Results to Hub

⚠️ EPHEMERAL ENVIRONMENT—MUST PUSH TO HUB

The Jobs environment is temporary. All files are deleted when the job ends. If the model isn't pushed to Hub, ALL TRAINING IS LOST.

Required Configuration

In training script/config:

SFTConfig(
    push_to_hub=True,
    hub_model_id="username/model-name",  # MUST specify
    hub_strategy="every_save",  # Optional: push checkpoints
)

In job submission:

{
    "secrets": {"HF_TOKEN": "$HF_TOKEN"}  # Enables authentication
}

Verification Checklist

Before submitting:

• [ ] push_to_hub=True set in config
• [ ] hub_model_id includes username/repo-name
• [ ] secrets parameter includes HF_TOKEN
• [ ] User has write access to target repo

See: references/hub_saving.md for detailed troubleshooting

Timeout Management

⚠️ DEFAULT: 30 MINUTES—TOO SHORT FOR TRAINING

Setting Timeouts

{
    "timeout": "2h"   # 2 hours (formats: "90m", "2h", "1.5h", or seconds as integer)
}

Timeout Guidelines

| Scenario | Recommended | Notes | |----------|-------------|-------| | Quick demo (50-100 examples) | 10-30 min | Verify setup | | Development training | 1-2 hours | Small datasets | | Production (3-7B model) | 4-6 hours | Full datasets | | Large model with LoRA | 3-6 hours | Depends on dataset |

Always add 20-30% buffer for model/dataset loading, checkpoint saving, Hub push operations, and network delays.

On timeout: Job killed immediately, all unsaved progress lost, must restart from beginning

Cost Estimation

Offer to estimate cost when planning jobs with known parameters. Use scripts/estimate_cost.py:

uv run scripts/estimate_cost.py \
  --model meta-llama/Llama-2-7b-hf \
  --dataset trl-lib/Capybara \
  --hardware a10g-large \
  --dataset-size 16000 \
  --epochs 3

Output includes estimated time, cost, recommended timeout (with buffer), and optimization suggestions.

When to offer: User planning a job, asks about cost/time, choosing hardware, job will run >1 hour or cost >$5

Example Training Scripts

Production-ready templates with all best practices:

Load these scripts for correctly:

• scripts/train_sft_example.py - Complete SFT training with Trackio, LoRA, checkpoints
• scripts/train_dpo_example.py - DPO training for preference learning
• scripts/train_grpo_example.py - GRPO training for online RL

These scripts demonstrate proper Hub saving, Trackio integration, checkpoint management, and optimized parameters. Pass their content inline to hf_jobs() or use as templates for custom scripts.

Monitorin

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: eugenepyvovarov
• Source: eugenepyvovarov/mcpbundler-agent-skills-marketplace
• License: MIT
• Homepage: https://mcp-bundler.com/skills-marketplace/

Install and usage instructions live in the source repository linked above.