Fine-Tuning Job Creation Workflow

This guide provides a step-by-step walkthrough for creating a fine-tuning job on Nexastack. Following these steps ensures your fine-tuning job is configured correctly and submitted successfully.

---

Goal

Learn how to:

• Create a new fine-tuning job for a selected base model.
• Upload training datasets and configure parameters.
• Review and submit jobs while verifying successful creation.

---

Step 1: Login to the Platform

1. Open the NexaStack login page.
2. Login with correct credientials

---

Step 2: Access Fine-Tuning Section

1. Navigate to the Marketplace or Fine-Tuning jobs from the sidebar.
2. Open the Fine-Tune Model section.
3. Click Create New Fine-Tuning Job.

---

Step 3: Select Base Model

1. Choose the base model you want to fine-tune.
2. Review the model details, capabilities, and limitations.

---

Step 4: Upload Training Dataset

1. Click Upload Training File.
2. Select the dataset file from your local system.
3. Confirm successful upload — a confirmation message or file preview should appear.

Dataset Requirements

• Ensure the dataset is in a supported format (e.g., JSON, CSV).
• Validate that the dataset contains the necessary fields for training.
• Large files may take longer to upload.

---

Step 5: Set Training Parameters

1. Enter a Job Name for identification.
2. Configure the training parameters:

Parameter	Description	Example
Epochs	Number of passes over the dataset	3–5
Batch Size	Number of samples per training batch	8–32
Learning Rate	Step size for optimization	5e-5–1e-4
Max Steps	Maximum number of training steps	Nearby 10
Model Length	Maximum token length per sample	Nearby 10

3. Validate all required fields are completed.

---

Step 6: Review and Submit

1. Review the training summary including dataset details and configured parameters.
2. Click Submit to start the fine-tuning job.

Pro Tip

Double-check all parameters before submitting. Incorrect settings can lead to suboptimal fine-tuning or job failure.

---

Step 7: Verify Job Creation

1. Check the backend/API to ensure the job is created successfully.
2. Confirm the new job appears in the list of fine-tuning jobs under the Fine-Tuning section.
3. Monitor the job status to track progress Pending → Running → Succeeded/Failed .

---

Best Practices

• Use descriptive job names for easy identification.
• Test your dataset with a small sample before full-scale training.
• Keep track of parameter combinations and results for reproducibility.
• Monitor system resources during fine-tuning to avoid interruptions.

---

Fine-Tuning Job Created

You have successfully created a fine-tuning job in Nexastack. The model is now queued for training and will be ready for deployment upon completion.