model_fine-tuning

Model Fine-Tuning

---

Quick Start

After signing in to the FPT AI Factory Portal at https://ai.fptcloud.com/ or
https://ai.fptcloud.jp/, follow the instruction below:

---

Step 1: Add credit to account

• Select Billing in the Sidemenu
• Click Add Credit button and enter an amount and payment information to complete.
  Or, if you have a voucher from FPT, apply your valid code in the Voucher code to redeem credits.
• Follow the detailed guide here:
  https://fptcloud.com/en/documents/ai-factory-billing/?doc=billing-calculation

Note:

• To start a fine-tuning job, you must maintain a minimum balance = number of nodes × number of GPUs × cost per block.
• If your balance falls below 0, the fine-tuning job will be canceled.
  After topping up, you will need to restart the job from the beginning.

---

Step 2: Create pipeline

• Select Model Fine-tuning in the Sidemenu
• Click the button Create New Pipeline and configure pipeline deployment.
• Follow the detailed guide here:
  https://fptcloud.com/en/documents/model-fine-tuning/?doc=create-pipeline

---

Step 3: Manage pipeline

• The pipeline list created successfully displays the Fine-tuning Pipelines page.
• In the Actions column, you can start, cancel, edit and delete pipelines to manage effectively.
• Click Pipeline name to access Execution history page, and click Execution name to access Execution details page.
• Access Execution history, navigate to tab Audit logs to view users and activities in the pipeline.
• Follow the detailed guide here:
  https://fptcloud.com/en/documents/model-fine-tuning/?doc=manage-pipeline

---

Step 4: Monitor and Evaluate

• Access Execution details, navigate to tab Model Metrics to monitor during the fine-tuning process and evaluate model after this process.
• Navigate to tab Logs to track fine-tuning processes, events, and system outputs in real time.
• Follow the detailed guide here:
  https://fptcloud.com/en/documents/model-fine-tuning/?doc=monitor-and-evaluate