gemma-7b-sft-qlora-no-robots15

This model is a fine-tuned version of google/gemma-7b on the chansung/no_robots_only_coding dataset. It achieves the following results on the evaluation set:

• Loss: 1.2808

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• total_eval_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: 15

Training results

Training Loss	Epoch	Step	Validation Loss
21.9058	0.91	5	7.6562
13.5645	2.0	11	6.6359
10.2613	2.91	16	6.0754
9.903	4.0	22	3.1116
4.594	4.91	27	1.6371
1.6122	6.0	33	1.4160
1.3971	6.91	38	1.3411
1.2757	8.0	44	1.3074
1.1233	8.91	49	1.2756
0.9741	10.0	55	1.2736
0.9266	10.91	60	1.2791
0.8584	12.0	66	1.2753
0.8714	12.91	71	1.2842
0.8421	13.64	75	1.2808

Framework versions

• PEFT 0.7.1
• Transformers 4.39.3
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2