Downloads · 30 days
16
37% of all-time downloads
bert_12_layer_model_v3_complete_training_new_emb_compress_48
gokuls/bert_12_layer_model_v3_complete_training_new_emb_compress_48
bert_12_layer_model_v3_complete_training_new_emb_compress_48 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. --
All-time downloads
43
Public
Repo size
10.6 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin407 MB · 100%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- Model type
- hybridbert
- Access
- Public
- Created
- Sep 26, 2023
- Updated
- Sep 29, 2023
- SHA
- 154896ab
Try a prompt
Model card
From the Hugging Face model README
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
bert_12_layer_model_v3_complete_training_new_emb_compress_48
This model is a fine-tuned version of on the gokuls/wiki_book_corpus_complete_processed_bert_dataset dataset. It achieves the following results on the evaluation set:
- Loss: 5.9594
- Accuracy: 0.1574
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: 48
- eval_batch_size: 48
- 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: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 7.1148 | 0.08 | 10000 | 7.0921 | 0.0828 |
| 6.6864 | 0.16 | 20000 | 6.6879 | 0.1078 |
| 6.5451 | 0.25 | 30000 | 6.5435 | 0.1184 |
| 6.4606 | 0.33 | 40000 | 6.4515 | 0.1262 |
| 6.3851 | 0.41 | 50000 | 6.3851 | 0.1312 |
| 6.3371 | 0.49 | 60000 | 6.3357 | 0.1342 |
| 6.2971 | 0.57 | 70000 | 6.2923 | 0.1373 |
| 6.2682 | 0.66 | 80000 | 6.2605 | 0.1399 |
| 6.2352 | 0.74 | 90000 | 6.2301 | 0.1411 |
| 6.214 | 0.82 | 100000 | 6.2090 | 0.1430 |
| 6.1837 | 0.9 | 110000 | 6.1865 | 0.1443 |
| 6.1726 | 0.98 | 120000 | 6.1682 | 0.1451 |
| 6.1524 | 1.07 | 130000 | 6.1498 | 0.1464 |
| 6.1293 | 1.15 | 140000 | 6.1300 | 0.1468 |
| 6.1116 | 1.23 | 150000 | 6.1026 | 0.1479 |
| 6.0839 | 1.31 | 160000 | 6.0797 | 0.1490 |
| 6.0616 | 1.39 | 170000 | 6.0590 | 0.1499 |
| 6.0508 | 1.47 | 180000 | 6.0399 | 0.1509 |
| 6.0311 | 1.56 | 190000 | 6.0233 | 0.1517 |
| 6.015 | 1.64 | 200000 | 6.0048 | 0.1533 |
| 5.985 | 1.72 | 210000 | 5.9863 | 0.1547 |
| 5.9661 | 1.8 | 220000 | 5.9595 | 0.1573 |
Framework versions
- Transformers 4.33.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.13.3
- Task
- Fill-Mask
- Library
- transformers
- Type
- hybridbert
- Created
- Sep 26, 2023
- Updated
- Sep 29, 2023