Gemma-2-2B_task-2_180-samples_config-1_full

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-2, the GaetanMichelet/chat-120_ft_task-2 and the GaetanMichelet/chat-180_ft_task-2 datasets. It achieves the following results on the evaluation set:

• Loss: 0.8685

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: 8
• total_train_batch_size: 8
• 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
1.2797	1.0	17	1.2698
1.1117	2.0	34	1.0882
0.9258	3.0	51	0.9472
0.8212	4.0	68	0.9018
0.7979	5.0	85	0.8801
0.7616	6.0	102	0.8685
0.7431	7.0	119	0.8698
0.6675	8.0	136	0.8850
0.5991	9.0	153	0.9122
0.5797	10.0	170	0.9582
0.559	11.0	187	1.0086
0.4197	12.0	204	1.1556
0.4433	13.0	221	1.1319

Framework versions

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