Downloads · 30 days
5
14% of all-time downloads
HBERTv1_emb_compress_48_L10_H128_A2
gokuls/HBERTv1_emb_compress_48_L10_H128_A2
HBERTv1_emb_compress_48_L10_H128_A2 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
36
Public
Repo size
899 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin25 MB · 96%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- Model type
- hybridbert
- Access
- Public
- Created
- Sep 27, 2023
- Updated
- Sep 28, 2023
- SHA
- 9a5f5277
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. -->
HBERTv1_emb_compress_48_L10_H128_A2
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: 6.2119
- Accuracy: 0.1437
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: 80
- eval_batch_size: 80
- 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 |
|---|---|---|---|---|
| 7.7198 | 0.14 | 10000 | 7.6446 | 0.0636 |
| 6.9424 | 0.27 | 20000 | 6.9379 | 0.0825 |
| 6.7559 | 0.41 | 30000 | 6.7492 | 0.0974 |
| 6.6513 | 0.55 | 40000 | 6.6455 | 0.1047 |
| 6.5706 | 0.68 | 50000 | 6.5734 | 0.1113 |
| 6.5236 | 0.82 | 60000 | 6.5176 | 0.1170 |
| 6.479 | 0.96 | 70000 | 6.4744 | 0.1212 |
| 6.4372 | 1.09 | 80000 | 6.4382 | 0.1255 |
| 6.4069 | 1.23 | 90000 | 6.4071 | 0.1275 |
| 6.3918 | 1.37 | 100000 | 6.3806 | 0.1302 |
| 6.3625 | 1.5 | 110000 | 6.3581 | 0.1320 |
| 6.339 | 1.64 | 120000 | 6.3400 | 0.1341 |
| 6.3189 | 1.78 | 130000 | 6.3237 | 0.1350 |
| 6.3115 | 1.91 | 140000 | 6.3103 | 0.1366 |
| 6.298 | 2.05 | 150000 | 6.2982 | 0.1374 |
| 6.2888 | 2.18 | 160000 | 6.2878 | 0.1382 |
| 6.2786 | 2.32 | 170000 | 6.2778 | 0.1392 |
| 6.2691 | 2.46 | 180000 | 6.2703 | 0.1398 |
| 6.2675 | 2.59 | 190000 | 6.2622 | 0.1402 |
| 6.255 | 2.73 | 200000 | 6.2571 | 0.1404 |
| 6.2481 | 2.87 | 210000 | 6.2517 | 0.1407 |
| 6.2465 | 3.0 | 220000 | 6.2447 | 0.1417 |
| 6.2381 | 3.14 | 230000 | 6.2399 | 0.1419 |
| 6.2332 | 3.28 | 240000 | 6.2371 | 0.1422 |
| 6.2279 | 3.41 | 250000 | 6.2321 | 0.1424 |
| 6.2275 | 3.55 | 260000 | 6.2274 | 0.1427 |
| 6.2186 | 3.69 | 270000 | 6.2249 | 0.1429 |
| 6.2192 | 3.82 | 280000 | 6.2231 | 0.1428 |
| 6.2228 | 3.96 | 290000 | 6.2208 | 0.1433 |
| 6.2154 | 4.1 | 300000 | 6.2192 | 0.1434 |
| 6.2204 | 4.23 | 310000 | 6.2177 | 0.1433 |
| 6.2123 | 4.37 | 320000 | 6.2162 | 0.1434 |
| 6.2136 | 4.51 | 330000 | 6.2134 | 0.1437 |
| 6.214 | 4.64 | 340000 | 6.2131 | 0.1439 |
| 6.2139 | 4.78 | 350000 | 6.2129 | 0.1437 |
| 6.211 | 4.92 | 360000 | 6.2130 | 0.1439 |
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 27, 2023
- Updated
- Sep 28, 2023