bert_12_layer_model_v1_complete_training_new_wt_init_48_frz

gokuls/bert_12_layer_model_v1_complete_training_new_wt_init_48_frz

bert_12_layer_model_v1_complete_training_new_wt_init_48_frz is a fill-mask model from gokuls. 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

29% of all-time downloads

All-time downloads

Public

Repo size

24.5 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin481 MB · 100%

At a glance

Task: Fill-Mask
Library: transformers
Model type: hybridbert
Access: Public
Created: Jun 20, 2023
Updated: Jun 23, 2023
SHA: ff1c3bdf

Try a prompt

Model card

From the Hugging Face model README

bert_12_layer_model_v1_complete_training_new_wt_init_48_frz

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

Loss: 2.2686
Accuracy: 0.5754

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: 1e-05
train_batch_size: 32
eval_batch_size: 32
seed: 10
distributed_type: multi-GPU
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 10000
num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
5.7497	0.05	10000	5.6455	0.1803
5.0738	0.11	20000	4.7886	0.2740
4.6823	0.16	30000	4.3914	0.3131
4.3942	0.22	40000	4.0874	0.3480
3.9186	0.27	50000	3.6561	0.3997
3.5977	0.33	60000	3.3709	0.4341
3.3917	0.38	70000	3.1996	0.4550
3.2539	0.44	80000	3.0836	0.4692
3.1642	0.49	90000	2.9976	0.4799
3.0815	0.55	100000	2.9278	0.4888
3.0111	0.6	110000	2.8682	0.4963
2.9572	0.66	120000	2.8223	0.5022
2.916	0.71	130000	2.7781	0.5077
2.8748	0.76	140000	2.7422	0.5121
2.8493	0.82	150000	2.7090	0.5167
2.8057	0.87	160000	2.6750	0.5207
2.7837	0.93	170000	2.6563	0.5236
2.7467	0.98	180000	2.6266	0.5271
2.7348	1.04	190000	2.6050	0.5303
2.7011	1.09	200000	2.5818	0.5330
2.6782	1.15	210000	2.5629	0.5356
2.658	1.2	220000	2.5464	0.5382
2.6375	1.26	230000	2.5260	0.5406
2.6257	1.31	240000	2.5065	0.5431
2.6122	1.37	250000	2.4934	0.5449
2.5836	1.42	260000	2.4741	0.5472
2.5706	1.47	270000	2.4647	0.5486
2.5536	1.53	280000	2.4519	0.5506
2.5403	1.58	290000	2.4376	0.5526
2.538	1.64	300000	2.4244	0.5538
2.5152	1.69	310000	2.4085	0.5558
2.4968	1.75	320000	2.4027	0.5576
2.4894	1.8	330000	2.3890	0.5586
2.4776	1.86	340000	2.3752	0.5606
2.476	1.91	350000	2.3631	0.5622
2.4541	1.97	360000	2.3559	0.5632
2.4474	2.02	370000	2.3448	0.5647
2.4434	2.08	380000	2.3349	0.5660
2.4324	2.13	390000	2.3300	0.5669
2.4264	2.18	400000	2.3182	0.5684
2.4034	2.24	410000	2.3082	0.5699
2.4034	2.29	420000	2.3007	0.5710
2.3946	2.35	430000	2.2905	0.5718
2.3816	2.4	440000	2.2830	0.5732
2.3774	2.46	450000	2.2771	0.5738
2.3651	2.51	460000	2.2686	0.5754

Framework versions

Transformers 4.30.2
Pytorch 1.14.0a0+410ce96
Datasets 2.13.0
Tokenizers 0.13.3

Task: Fill-Mask
Library: transformers
Type: hybridbert
Created: Jun 20, 2023
Updated: Jun 23, 2023