train_run-qwen2.5-0.5b-instruct-arabic-diacritization-fadel-10k-left-3e

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B-Instruct on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0889

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.PAGED_ADAMW with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
0.2681	0.16	100	0.2446
0.176	0.32	200	0.1755
0.15	0.48	300	0.1536
0.1418	0.64	400	0.1301
0.1225	0.8	500	0.1217
0.1104	0.96	600	0.1114
0.0804	1.12	700	0.1133
0.0833	1.28	800	0.1018
0.0796	1.44	900	0.1016
0.0812	1.6	1000	0.0986
0.0874	1.76	1100	0.0966
0.0738	1.92	1200	0.0912
0.0521	2.08	1300	0.0951
0.0508	2.24	1400	0.0918
0.0512	2.4	1500	0.0898
0.054	2.56	1600	0.0933
0.053	2.7200	1700	0.0928
0.0562	2.88	1800	0.0889

Framework versions

• PEFT 0.14.0
• Transformers 4.51.1
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.0