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

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

• Loss: 0.4006

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
3.072	0.6957	2	3.0875
3.0765	1.7391	5	2.5635
2.0933	2.7826	8	1.4153
1.2307	3.8261	11	0.7854
0.5656	4.8696	14	0.5590
0.5325	5.9130	17	0.4852
0.4027	6.9565	20	0.4474
0.2861	8.0	23	0.4358
0.3766	8.6957	25	0.4268
0.291	9.7391	28	0.4162
0.2706	10.7826	31	0.4064
0.2037	11.8261	34	0.4027
0.2057	12.8696	37	0.4006
0.1402	13.9130	40	0.4081
0.1817	14.9565	43	0.4246
0.119	16.0	46	0.4514
0.1067	16.6957	48	0.4775
0.0954	17.7391	51	0.5345
0.0553	18.7826	54	0.5770
0.0604	19.8261	57	0.6300

Framework versions

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