Korla/hsb-mistral

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

hsb-mistral-7b

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.2.

Model description

This is a fine-tune for the upper sorbian language.

Intended uses & limitations

This model is merely an experiment and simply a plaything; it may generate inaccurate results.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• —learning_rate: 6.25e-05
• —trainbatchsize: 1
• —evalbatchsize: 8
• —seed: 42
• —gradientaccumulationsteps: 8
• —totaltrainbatch_size: 8
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: cosine
• —lrschedulerwarmup_ratio: 0.05
• —num_epochs: 3

Training results

Framework versions

• —Transformers 4.39.0.dev0
• —Pytorch 2.2.0
• —Datasets 2.17.1
• —Tokenizers 0.15.2