results_ara_t5_v2

This model is a fine-tuned version of bakrianoo/t5-arabic-base on an unknown dataset. It achieves the following results on the evaluation set:

• Gen Len: 19.0
• Loss: 0.0001
• Rouge1: 0.0875
• Rouge2: 0.0046
• Rougel: 0.0871
• Rougelsum: 0.0873

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: 5e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Gen Len	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum
5.4255	0.43	500	18.8641	0.0292	0.0123	0.0028	0.0122	0.0122
0.0975	0.86	1000	18.7006	0.0101	0.0146	0.0038	0.0146	0.0146
0.0321	1.29	1500	18.5583	0.0077	0.078	0.0046	0.0779	0.078
0.0237	1.71	2000	18.7994	0.0075	0.0806	0.0066	0.0809	0.0805
0.0182	2.14	2500	19.0	0.0072	0.0761	0.0023	0.0759	0.076
0.0148	2.57	3000	18.8528	0.0069	0.0952	0.0061	0.0957	0.0951
0.0156	3.0	3500	19.0	0.0064	0.0773	0.0036	0.0772	0.0772
0.0126	3.43	4000	18.5583	0.0060	0.0859	0.0038	0.0862	0.0862
0.0165	3.86	4500	18.8528	0.0056	0.0832	0.0075	0.0835	0.0833
0.0101	4.28	5000	19.0	0.0047	0.0846	0.0052	0.0853	0.0848
0.0089	4.71	5500	19.0	0.0022	0.0874	0.0046	0.0871	0.0873
0.0063	5.14	6000	19.0	0.0009	0.0875	0.0046	0.0871	0.0873
0.0051	5.57	6500	19.0	0.0003	0.0875	0.0046	0.0871	0.0873
0.0042	6.0	7000	19.0	0.0002	0.0875	0.0046	0.0871	0.0873
0.0037	6.43	7500	19.0	0.0002	0.0875	0.0046	0.0871	0.0873
0.0035	6.86	8000	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.003	7.28	8500	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.0027	7.71	9000	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.0024	8.14	9500	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.0023	8.57	10000	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.0022	9.0	10500	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.002	9.43	11000	19.0	0.0001	0.0875	0.0046	0.0871	0.0873
0.0019	9.85	11500	19.0	0.0001	0.0875	0.0046	0.0871	0.0873

Framework versions

• Transformers 4.40.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2