Gemma-2-2B_task-3_60-samples_config-2

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4126

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
3.2454	0.6957	2	3.2345
3.2274	1.7391	5	2.6911
2.1741	2.7826	8	1.4943
1.3679	3.8261	11	0.8702
0.6621	4.8696	14	0.6007
0.5808	5.9130	17	0.5171
0.456	6.9565	20	0.4745
0.3027	8.0	23	0.4504
0.3998	8.6957	25	0.4361
0.2969	9.7391	28	0.4208
0.283	10.7826	31	0.4126
0.1909	11.8261	34	0.4143
0.1833	12.8696	37	0.4323
0.1128	13.9130	40	0.4549
0.16	14.9565	43	0.4737
0.0817	16.0	46	0.5361
0.0679	16.6957	48	0.5978
0.0501	17.7391	51	0.6335

Framework versions

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