Gemma-2-2B_task-1_60-samples_config-1

This model is a fine-tuned version of google/gemma-2-2b-it on the GaetanMichelet/chat-60_ft_task-1 dataset. It achieves the following results on the evaluation set:

• Loss: 1.3620

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
2.489	0.8696	5	2.2686
2.1234	1.9130	11	1.8447
1.5886	2.9565	17	1.5018
1.4359	4.0	23	1.3991
1.1595	4.8696	28	1.3782
0.9043	5.9130	34	1.3620
0.773	6.9565	40	1.4541
0.4933	8.0	46	1.6163
0.354	8.8696	51	1.9201
0.1478	9.9130	57	2.4809
0.0676	10.9565	63	2.9184
0.0308	12.0	69	3.2972
0.0342	12.8696	74	3.3674

Framework versions

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