miti-2.3b-base-v1

Beta — first Kyrgyz-language large language model built from scratch Still experimental in coherence, the model produces somewhat grammatically correct and fluent sentences, contributing to low-resource NLP research for the Kyrgyz language.

P.S: I don't have enough data to make a better pre-train so project is on hold for now.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 1

(technically run several epochs not 1)

Training results

Framework versions

• Transformers 4.56.2
• Pytorch 2.8.0.dev20250319+cu128
• Datasets 4.1.1
• Tokenizers 0.22.1

Metinov Adilet 2025