Training as a Service (TaaS)

With IO Intelligence’s Training as a Service (TaaS), you can fine-tune powerful pre-trained models using your own data—achieving tailored performance without the complexity of training from scratch. Built for builders, researchers, developers, and businesses, TaaS gives you greater control over how your AI learns and adapts. With support for advanced techniques such as PPO, DPO, and reward modeling, you can move beyond traditional fine-tuning and experiment at the cutting edge of machine learning.

​

Getting Started

To begin training your model:

• Go to your io.net Dashboard.
• Navigate to Training under the IO Intelligence section.
  
• Click the Start Training button to launch the setup form.
  

​

Build Your Training Workflow

The Training Model form is designed to guide you through the setup process, step by step.

​

1. Select Training Method

Pick how your model should learn. You can choose from a variety of advanced methods:

• Supervised Fine-Tuning: Teach your model using labeled datasets for task-specific learning.
• Reward Modeling: Train your model to assign scores to generated responses.
• Proximal Policy Optimization (PPO): Use reinforcement learning with reward feedback.
• Direct Preference Optimization (DPO): Optimize the model directly using ranked preferences.
• Controlled Tuning Optimization (Experimental): Apply KL-regularized tuning for fine-grained control.

​

2. Select a Base Model

You can select a pre-integrated model from our library or bring your own model from Hugging Face. Option A - Use a Preloaded Model: Choose from popular base models in the dropdown list - e.g., LLaMA, Mistral, Falcon, GPT-Neo, etc. Option B - Link a Hugging Face Model:

1. Paste your Hugging Face model repo link in the field.
2. Click the Test button.
3. If the connection is successful, you will see: “Successfully tested.”
4. If the repository is private or invalid, a prompt will appear requesting your Hugging Face Token, with the following message: “The repo is private or does not exist. Provide an HF Token.” For more information, refer to How to get your Hugging Face Token.

​

3. Select a Dataset

Choose training data from a curated list or bring your own custom dataset. Option A - Use a Built-in Dataset: Select from our list of available datasets using the dropdown. Option B - Link a Custom Dataset: Set up your dataset by completing the configuration fields.

​

4. Choose a Training Style

Depending on your goal, you can select between Simple Creating and Advanced Creating when training a model. Option A - Simple Creating: Uses default settings for a quick and straightforward setup. Best if you want to get started fast without managing technical details. Option B - Advanced Creating: Offers full flexibility and control. Choose this if your goal is to customize every aspect of how your model learns, from hyperparameters to optimization strategies.

​

5. Start Training

After completing all fields, click Start Training to begin.

​

Training Dashboard

The following sections can be viewed in the Training Dashboard.

​

Your Current Plan

At the top of the page, you can view your current plan details. To increase your training runs or access faster processing, click the Upgrade button.

​

Your Training Jobs

Below your plan, there is a Training Jobs table where you can view and manage all your model training requests. Each row shows key details such as, Job ID, Base Model, Type, Status, and Run Time. Click on any job to open the detailed view and see how it is progressing.

​

Job Details

Click on a job from the dashboard to open its Job Details. This screen provides everything you need to track and manage your model training in real time. At the top of the page are the following options and indicators:

• Download Model – Available once the job is complete, to download the final model.
• Abort Training – Manually stop the job if necessary.
• Time Remaining – Displays how much time is left for training to finish.
• Time Passed – Shows how long the job has been running.

​

Training Metrics (Charts)

Track your model’s learning performance in real time.

• Loss Chart – Shows how training loss decreases over time.

This visual tool allows you to understand if your model is learning effectively - or if it needs adjustments.

​

Training Details

A summary of the key information for the training job is provided as follows:

• Status – Created, Deploying, Deploy Failed, Training Failed, Training, or Completed.
• Date Created
• Model Name
• Training Method
• Base Model Used
• Dataset Used
• User Tag - Custom label or identifier.
• End Date - If completed or aborted.

​

Training Logs

Training Logs provide visibility into what occurs behind the scenes during the model training process. They contain a detailed list of steps, events, and status updates throughout the job’s lifecycle, making them especially useful for debugging, monitoring, or maintaining transparency. To download the logs, select the Download Logs button to save them locally for review or record-keeping.