TindeBielik_finetunes

This model is a fine-tuned version of speakleash/Bielik-7B-Instruct-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5315

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: 8e-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 0.03
• training_steps: 75

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.2632	5	2.2403
2.2352	0.5263	10	1.9857
2.2352	0.7895	15	1.8578
1.8626	1.0526	20	1.7444
1.8626	1.3158	25	1.6693
1.5773	1.5789	30	1.6168
1.5773	1.8421	35	1.5745
1.4663	2.1053	40	1.5478
1.4663	2.3684	45	1.5356
1.1649	2.6316	50	1.5239
1.1649	2.8947	55	1.5185
0.9855	3.1579	60	1.5044
0.9855	3.4211	65	1.5510
0.6071	3.6842	70	1.5145
0.6071	3.9474	75	1.5315

Framework versions

• PEFT 0.10.0
• Transformers 4.40.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1