README.md

metadata

tags:
  - generated_from_trainer
metrics:
  - bleu
model-index:
  - name: BERT2BERT_pretrained_LC-QuAD_2.0
    results: []

BERT2BERT_pretrained_LC-QuAD_2.0

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9337
• Bleu: 81.2301
• Em: 0.1
• Rm: 0.22
• Gen Len: 46.9

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2500
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Em	Rm	Gen Len
8.0718	0.83	2000	6.9689	5.8841	0.0	0.0	126.98
3.6551	1.66	4000	3.5248	20.8847	0.0	0.0	30.46
2.3528	2.49	6000	2.1427	47.9723	0.0	0.02	44.52
1.5534	3.32	8000	1.5217	63.3606	0.0	0.06	46.32
1.1024	4.15	10000	1.2364	67.3193	0.04	0.14	42.5
0.9597	4.98	12000	0.9974	74.2575	0.04	0.16	47.52
0.7033	5.81	14000	0.9390	74.9164	0.08	0.2	45.64
0.5446	6.64	16000	0.9508	75.9224	0.06	0.14	47.34
0.4362	7.47	18000	0.8897	77.2209	0.1	0.22	44.36
0.3231	8.3	20000	0.9232	78.5526	0.16	0.26	46.66
0.2584	9.13	22000	0.9404	79.9787	0.1	0.2	47.52
0.2387	9.96	24000	0.9337	81.2301	0.1	0.22	46.9

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3