mt5-small_test_35

This model is a fine-tuned version of google/mt5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7383
• Rouge1: 43.9482
• Rouge2: 38.4156
• Rougel: 42.6232
• Rougelsum: 42.674
• Bleu: 33.3469
• Gen Len: 12.4725
• Meteor: 0.4016
• True negatives: 70.997
• False negatives: 11.8271
• Cosine Sim: 0.7532

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: 0.001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 9
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bleu	Gen Len	Meteor	True negatives	False negatives	Cosine Sim
2.4524	1.0	175	0.9783	17.6419	14.587	17.1176	17.1329	6.1296	7.3271	0.1531	75.7704	59.8602	0.3786
1.1433	1.99	350	0.8448	38.9957	33.2414	37.7868	37.8653	27.5883	12.3274	0.3526	60.3625	17.236	0.6954
0.9381	2.99	525	0.8067	42.4146	36.3126	40.964	41.0427	31.5838	13.0716	0.3833	59.6375	11.1801	0.7425
0.8116	3.98	700	0.7712	43.8741	37.8446	42.3785	42.4778	33.1873	13.0574	0.3982	61.9335	9.5238	0.7586
0.7218	4.98	875	0.7439	43.1579	37.3057	41.7059	41.8024	32.5124	12.7853	0.3931	65.8006	11.2836	0.7498
0.6461	5.97	1050	0.7254	39.9226	34.552	38.7033	38.7665	27.9936	11.4675	0.3638	77.9456	18.5041	0.7003
0.5852	6.97	1225	0.7290	44.131	38.3527	42.7974	42.8549	33.6955	12.7811	0.4026	67.855	10.3778	0.7599
0.5421	7.96	1400	0.7248	44.5368	38.7443	43.2111	43.2976	34.1121	12.7875	0.4071	67.5529	10.4037	0.7637
0.5026	8.96	1575	0.7383	43.9482	38.4156	42.6232	42.674	33.3469	12.4725	0.4016	70.997	11.8271	0.7532

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3