Gemma-2-2B_task-3_180-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, the GaetanMichelet/chat-120_ft_task-3_auto and the GaetanMichelet/chat-180_ft_task-3_auto datasets. It achieves the following results on the evaluation set:

• Loss: 0.3731

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
3.1214	0.9412	8	2.8617
1.4487	2.0	17	1.1770
0.6083	2.9412	25	0.5479
0.3481	4.0	34	0.4357
0.3365	4.9412	42	0.3944
0.3762	6.0	51	0.3785
0.2475	6.9412	59	0.3731
0.2295	8.0	68	0.3856
0.1897	8.9412	76	0.4363
0.1273	10.0	85	0.5127
0.1212	10.9412	93	0.5698
0.031	12.0	102	0.7057
0.0283	12.9412	110	0.7808
0.0227	14.0	119	0.8176

Framework versions

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