PipeFactory
v0.1.0
The dataset will be streamed from HuggingFace at training time. Make sure the dataset is public or your HF token has access.
For instruction-tuning (qlora/lora/full_sft), your dataset should contain instruction and output fields, or be in ShareGPT/ChatML format.
Defaults pre-filled from LoRA preset. Adjust as needed.
Estimated cost
$1.50/hr
Hardware
A10G Large (24GB)
Push to HuggingFace Hub
Automatically upload the trained adapter to your HF account after training completes.
Review your pipeline configuration before saving.