Modelpublic

NotASI/Gemma-2-9B-Alpaca-QDoRA-step120

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face

0likes10downloads

Model Card

Uploaded model

—Developed by: NotAiLOL
—License: apache-2.0
—Finetuned from model : unsloth/gemma-2-9b-bnb-4bit

This gemma2 model was trained 2x faster with Unsloth and Huggingface's TRL library.

Details

This model is fine tuned from unsloth/gemma-2-9b-bnb-4bit on the alpaca-cleaned dataset using the QDoRA method.

This model achieved a loss of 0.817200 on the alpaca-cleaned dataset after step 120.

This model follows the alpaca prompt:

Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
{}

### Input:
{}

### Response:
{}

Training

This model is trained on a single Tesla T4 GPU.

—2047.6215 seconds used for training.
—34.13 minutes used for training.
—Peak reserved memory = 9.562 GB.
—Peak reserved memory for training = 2.216 GB.
—Peak reserved memory % of max memory = 64.836 %.
—Peak reserved memory for training % of max memory = 15.026 %.