bert_character

RafaelEiji/bert_character

bert_character is a fill-mask model from RafaelEiji. Use it when you need the model to fill a missing word. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

1% of all-time downloads

All-time downloads

794

Public

Repo size

1.4 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin345 MB · 100%

At a glance

Task: Fill-Mask
Library: transformers
Model type: bert
Access: Public
Created: Jan 31, 2023
Updated: Mar 22, 2023
SHA: 44bd42fc

Try a prompt

Model card

From the Hugging Face model README

from_scratch

This model is a fine-tuned version of tokenizer/config.json on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4744

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: 0.0001
train_batch_size: 360
eval_batch_size: 360
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 10000
num_epochs: 3.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
1.0952	0.05	20000	1.0383
0.936	0.1	40000	0.8852
0.8679	0.14	60000	0.8207
0.8276	0.19	80000	0.7796
0.796	0.24	100000	0.7519
0.7756	0.29	120000	0.7299
0.7545	0.33	140000	0.7103
0.7395	0.38	160000	0.6947
0.7236	0.43	180000	0.6809
0.7143	0.48	200000	0.6705
0.705	0.52	220000	0.6585
0.6904	0.57	240000	0.6479
0.6835	0.62	260000	0.6388
0.672	0.67	280000	0.6290
0.665	0.72	300000	0.6217
0.6581	0.76	320000	0.6136
0.6466	0.81	340000	0.6071
0.6396	0.86	360000	0.6000
0.6343	0.91	380000	0.5940
0.6286	0.95	400000	0.5880
0.6183	1.0	420000	0.5809
0.6134	1.05	440000	0.5757
0.6094	1.1	460000	0.5693
0.6032	1.15	480000	0.5641
0.5954	1.19	500000	0.5596
0.5915	1.24	520000	0.5532
0.5845	1.29	540000	0.5489
0.5823	1.34	560000	0.5437
0.5754	1.38	580000	0.5393
0.573	1.43	600000	0.5345
0.5643	1.48	620000	0.5309
0.5627	1.53	640000	0.5262
0.56	1.57	660000	0.5220
0.5554	1.62	680000	0.5186
0.5507	1.67	700000	0.5152
0.5494	1.72	720000	0.5117
0.5445	1.77	740000	0.5076
0.5396	1.81	760000	0.5051
0.5363	1.86	780000	0.5026
0.5356	1.91	800000	0.4998
0.5303	1.96	820000	0.4982
0.5583	2.0	840000	0.5195
0.5565	2.05	860000	0.5180
0.5535	2.1	880000	0.5158
0.5497	2.15	900000	0.5133
0.5511	2.19	920000	0.5110
0.5439	2.24	940000	0.5085
0.5413	2.29	960000	0.5060
0.5376	2.34	980000	0.5023
0.5333	2.39	1000000	0.5004
0.5322	2.43	1020000	0.4973
0.5312	2.48	1040000	0.4941
0.5281	2.53	1060000	0.4921
0.5267	2.58	1080000	0.4902
0.5257	2.62	1100000	0.4871
0.5174	2.67	1120000	0.4849
0.5183	2.72	1140000	0.4825
0.5181	2.77	1160000	0.4807
0.5116	2.81	1180000	0.4784
0.5092	2.86	1200000	0.4769
0.5109	2.91	1220000	0.4757
0.5102	2.96	1240000	0.4739

Framework versions

Transformers 4.25.1
Pytorch 1.13.1+cu117
Datasets 2.8.0
Tokenizers 0.13.2

Task: Fill-Mask
Library: transformers
Type: bert
Created: Jan 31, 2023
Updated: Mar 22, 2023