Gemma-2-2B_task-3_120-samples_config-2_full

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-3 and the GaetanMichelet/chat-120_ft_task-3 datasets. It achieves the following results on the evaluation set:

• Loss: 0.9439

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
1.4237	0.9091	5	1.3857
1.3836	2.0	11	1.2989
1.2209	2.9091	16	1.2033
1.129	4.0	22	1.1133
1.0353	4.9091	27	1.0336
0.957	6.0	33	0.9890
0.9251	6.9091	38	0.9710
0.9012	8.0	44	0.9577
0.8677	8.9091	49	0.9495
0.8692	10.0	55	0.9446
0.8262	10.9091	60	0.9439
0.7985	12.0	66	0.9476
0.7634	12.9091	71	0.9547
0.7465	14.0	77	0.9688
0.6551	14.9091	82	0.9894
0.6593	16.0	88	1.0191
0.599	16.9091	93	1.0529
0.5401	18.0	99	1.0786

Framework versions

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