llama32-1b-fineweb-is-lora-peft-effbatch512

This model is a fine-tuned version of meta-llama/Llama-3.2-1B-Instruct on the Icelandic portion of Fineweb-2.

Model description

LoRA rank 256, alpha 512. Target modules: ["up_proj", "down_proj", "o_proj"]

Intended uses & limitations

More information needed

Training and evaluation data

Training data: Icelandic portion of Fineweb-2

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 1024
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 1

Training results

Framework versions

• PEFT 0.13.2
• Transformers 4.43.4
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1