G0514HMA17H

This model is a fine-tuned version of google/gemma-2b on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: -17.8863

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_steps: 60
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.7397	0.09	10	-0.3975
-1.7057	0.18	20	-3.6259
-5.262	0.27	30	-7.4840
-9.0284	0.36	40	-11.0077
-12.2555	0.45	50	-13.9478
-14.8695	0.54	60	-15.9731
-16.4824	0.63	70	-16.9638
-17.1562	0.73	80	-17.3577
-17.4316	0.82	90	-17.5211
-17.562	0.91	100	-17.6057
-17.6246	1.0	110	-17.6531
-17.6704	1.09	120	-17.6811
-17.691	1.18	130	-17.7006
-17.7076	1.27	140	-17.7180
-17.7234	1.36	150	-17.7314
-17.738	1.45	160	-17.7406
-17.7492	1.54	170	-17.7580
-17.7665	1.63	180	-17.7743
-17.7825	1.72	190	-17.7921
-17.8068	1.81	200	-17.8132
-17.8209	1.9	210	-17.8311
-17.8385	1.99	220	-17.8439
-17.8464	2.08	230	-17.8522
-17.8595	2.18	240	-17.8657
-17.8688	2.27	250	-17.8729
-17.8733	2.36	260	-17.8783
-17.8725	2.45	270	-17.8809
-17.8806	2.54	280	-17.8834
-17.8837	2.63	290	-17.8849
-17.884	2.72	300	-17.8855
-17.8807	2.81	310	-17.8860
-17.8864	2.9	320	-17.8863
-17.8839	2.99	330	-17.8863

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.2+cu121
• Datasets 2.14.6
• Tokenizers 0.14.0