Gemma-2-2B_task-2_60-samples_config-2

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6932

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 16
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss
1.0008	0.6957	2	1.0610
0.9848	1.7391	5	1.0090
0.8909	2.7826	8	0.8771
0.7601	3.8261	11	0.8055
0.7251	4.8696	14	0.7676
0.6518	5.9130	17	0.7352
0.6256	6.9565	20	0.7191
0.5985	8.0	23	0.7080
0.5809	8.6957	25	0.7035
0.5221	9.7391	28	0.6958
0.5052	10.7826	31	0.6939
0.4727	11.8261	34	0.6932
0.4384	12.8696	37	0.6968
0.4074	13.9130	40	0.7064
0.3396	14.9565	43	0.7210
0.3296	16.0	46	0.7495
0.2865	16.6957	48	0.7752
0.2732	17.7391	51	0.8188
0.2207	18.7826	54	0.8808

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.1.2+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1