poziomka_11_sft_lora_2026_03_24

<!-- 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. -->

poziomka-11-sft-lora-2026-03-24-1

This model is a fine-tuned version of cpral/poziomka_11_hf/iter_0006000 on the poziomka-sft-2 dataset. It achieves the following results on the evaluation set:

• Loss: 1.5421

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: 0.0003
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 16
• total_train_batch_size: 256
• total_eval_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 40
• num_epochs: 1.0

Training results

Framework versions

• PEFT 0.18.1
• Transformers 4.55.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4