isekai-bert-v1

isek-ai/isekai-bert-v1

isekai-bert-v1 is a fill-mask model from isek-ai. Use it when you need the model to fill a missing word. It is set up for transformers. The card lists the license as apache-2.0.

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

Downloads · 30 days

5% of all-time downloads

All-time downloads

417

Public

Parameters

111M

6.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors223 MB · 100%

At a glance

Task: Fill-Mask
Library: transformers
License: apache-2.0
Model type: bert
Access: Public
Created: Oct 17, 2023
Updated: Oct 17, 2023
SHA: db6575ed

Try a prompt

Base models

Model card

From the Hugging Face model README

isekai-bert-v1

This model is a fine-tuned version of cl-tohoku/bert-base-japanese-v3 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.9164

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: 0.0005
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: cosine
lr_scheduler_warmup_steps: 500
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
2.9144	0.06	1000	2.6322
3.0793	0.13	2000	2.7752
3.0591	0.19	3000	2.8336
2.972	0.26	4000	2.9084
2.9967	0.32	5000	2.8845
2.9489	0.38	6000	2.7112
2.8639	0.45	7000	2.7209
2.8355	0.51	8000	2.6684
2.8162	0.58	9000	2.6209
2.7648	0.64	10000	2.5749
2.6663	0.7	11000	2.5231
2.6451	0.77	12000	2.4754
2.6041	0.83	13000	2.4279
2.5306	0.9	14000	2.3829
2.4765	0.96	15000	2.3137
2.3899	1.02	16000	2.3052
2.3681	1.09	17000	2.2123
2.2821	1.15	18000	2.1934
2.2288	1.22	19000	2.1399
2.1858	1.28	20000	2.0922
2.1964	1.34	21000	2.0689
2.1419	1.41	22000	2.0357
2.1011	1.47	23000	2.0327
2.039	1.54	24000	1.9853
2.0284	1.6	25000	1.9778
2.0253	1.66	26000	1.9869
2.0292	1.73	27000	1.9494
2.0016	1.79	28000	1.9158
2.0387	1.86	29000	1.9778
1.9679	1.92	30000	1.9171
2.0441	1.98	31000	1.9164

Framework versions

Transformers 4.34.0
Pytorch 2.0.0+cu118
Datasets 2.14.5
Tokenizers 0.14.0

Task: Fill-Mask
Library: transformers
Type: bert
License: apache-2.0
Languages: ja
Created: Oct 17, 2023
Updated: Oct 17, 2023