Modelpublic

DooDooHyun/AIFT-1.3B-instruct-all-v0.54

sourceHugging Facecc-by-nc-4.0updated 3y agoView on Hugging Face

0likes17downloads

Model Card

AIFT-1.3B-instruct-all-v0.54

Model description

This model is a fine-tuned version of 42dot/42dot_LLM-PLM-1.3B.

Intended uses & limitations

More information needed

Training data

More information needed

Training procedure

Hardware

—ubuntu 18.04
—GPU : A6000 * 3ea

Training hyperparameters

The following hyperparameters were used during training:

—learning_rate: 2e-05
—trainbatchsize: 24
—evalbatchsize: 8
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 96
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 1.0
—mixedprecisiontraining: Native AMP

Model Benchmark

Evaluation

ko-lm-eval-harness

Ko-LLM leaderboard

—Follow up as Open KO-LLM LeaderBoard.

Framework versions

—Transformers 4.36.2
—Pytorch 2.1.2+cu121
—Datasets 2.0.0
—Tokenizers 0.15.0

License

The model is licensed under the cc-by-nc-sa-4.0 license, which allows others to copy, modify, and share the work non-commercially, as long as they give appropriate credit and distribute any derivative works under the same license.