train_cornel_obj_cluster_hpoparams

This model is a fine-tuned version of google/gemma-2b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.7681
• Rouge1: 0.2969
• Rouge2: 0.0809
• Rougel: 0.2831
• Rougelsum: 0.2829

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: 0.00021
• train_batch_size: 20
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 40
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 2
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
4.1083	0.2041	25	4.0315	0.2893	0.0769	0.2705	0.2691
3.9688	0.4082	50	3.8554	0.2921	0.0838	0.2751	0.2744
3.9586	0.6122	75	3.7972	0.2974	0.0821	0.2805	0.2807
3.6753	0.8163	100	3.7705	0.3023	0.0805	0.2826	0.2823
3.5398	1.0204	125	3.7593	0.3050	0.0793	0.2855	0.2848
3.5285	1.2245	150	3.7672	0.3040	0.0790	0.2871	0.2870
3.6062	1.4286	175	3.7700	0.3005	0.0797	0.2850	0.2847
3.3567	1.6327	200	3.7655	0.2995	0.0827	0.2854	0.2853
3.5132	1.8367	225	3.7681	0.2969	0.0809	0.2831	0.2829

Framework versions

• PEFT 0.11.1
• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1