Fine-tune - SPUR Compute

Fine-tune

Fine-tune an open model on Canadian iron.

Upload your data, fine-tune an open model on SPUR-hosted NVIDIA Blackwell in Waterloo, and your private model is served right on your account - callable from the same API as every other model. Your data stays in Canada.

$0.00

Current credit balance. Each fine-tune debits this balance when it starts.

Add credit

Create a fine-tune

Base model

Pick the open model to tune. The estimate is GPU-hours for one epoch.

Loading base models...

Name

Training data

Format: JSONL - one example per line, chat messages. Minimum 10 examples, up to 200MB.

Drop a .jsonl file here, or click to choose

Each line: a chat conversation.

{"messages":[{"role":"user","content":"Hi"},{"role":"assistant","content":"Hello! How can I help?"}]}
{"prompt":"Translate: cat","completion":"chat"}

Epochs

LoRA rank

Learning rate

$0.00

Select a base model to estimate cost.

$0.00

your balance

Fine-tune an open model on Canadian iron.

Create a fine-tune

Your fine-tunes