gemma2b-closedqa-gpt4o-100k

This model is a fine-tuned version of google/gemma-2b on the llama-duo/synth_closed_qa_dataset_dedup dataset. It achieves the following results on the evaluation set:

• Loss: 2.6811

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 3
• gradient_accumulation_steps: 2
• total_train_batch_size: 48
• total_eval_batch_size: 24
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss
1.0659	0.9983	287	2.2446
0.9881	2.0	575	2.2538
0.9259	2.9983	862	2.2906
0.8772	4.0	1150	2.3648
0.8436	4.9983	1437	2.4373
0.7905	6.0	1725	2.5194
0.757	6.9983	2012	2.6101
0.7274	8.0	2300	2.6532
0.721	8.9983	2587	2.6792
0.7142	9.9826	2870	2.6811

Framework versions

• PEFT 0.11.1
• Transformers 4.41.2
• Pytorch 2.2.2+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1