Downloads · 30 days
6
3% of all-time downloads
bert_12_layer_model_v1_complete_training
gokuls/bert_12_layer_model_v1_complete_training
bert_12_layer_model_v1_complete_training 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
200
Public
Repo size
26.3 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin212 MB · 99%
At a glance
- Task
- Fill-Mask
- Library
- transformers
- Model type
- hybridbert
- Access
- Public
- Created
- Feb 28, 2023
- Updated
- Mar 6, 2023
- SHA
- 0fa47701
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_v1_complete_training
This model is a fine-tuned version of on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.7148
- Accuracy: 0.6576
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: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- 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 |
|---|---|---|---|---|
| 6.1779 | 0.11 | 10000 | 6.1719 | 0.1487 |
| 4.6914 | 0.22 | 20000 | 4.5039 | 0.3178 |
| 3.2325 | 0.33 | 30000 | 3.0977 | 0.4772 |
| 2.831 | 0.44 | 40000 | 2.7266 | 0.5224 |
| 2.6262 | 0.55 | 50000 | 2.5371 | 0.5455 |
| 2.5006 | 0.66 | 60000 | 2.4141 | 0.5614 |
| 2.4062 | 0.76 | 70000 | 2.3242 | 0.5734 |
| 2.3338 | 0.87 | 80000 | 2.2539 | 0.5823 |
| 2.2838 | 0.98 | 90000 | 2.2012 | 0.5894 |
| 2.231 | 1.09 | 100000 | 2.1504 | 0.5959 |
| 2.1903 | 1.2 | 110000 | 2.1133 | 0.6009 |
| 2.1594 | 1.31 | 120000 | 2.0801 | 0.6054 |
| 2.1307 | 1.42 | 130000 | 2.0488 | 0.6095 |
| 2.0948 | 1.53 | 140000 | 2.0234 | 0.6133 |
| 2.0748 | 1.64 | 150000 | 1.9980 | 0.6169 |
| 2.0572 | 1.75 | 160000 | 1.9756 | 0.6195 |
| 2.0359 | 1.86 | 170000 | 1.9551 | 0.6225 |
| 2.0148 | 1.97 | 180000 | 1.9385 | 0.6251 |
| 1.9994 | 2.08 | 190000 | 1.9219 | 0.6274 |
| 1.9769 | 2.18 | 200000 | 1.9043 | 0.6297 |
| 1.9705 | 2.29 | 210000 | 1.8916 | 0.6317 |
| 1.9557 | 2.4 | 220000 | 1.8779 | 0.6338 |
| 1.9407 | 2.51 | 230000 | 1.8643 | 0.6354 |
| 1.9307 | 2.62 | 240000 | 1.8525 | 0.6372 |
| 1.9186 | 2.73 | 250000 | 1.8408 | 0.6388 |
| 1.9114 | 2.84 | 260000 | 1.8320 | 0.6401 |
| 1.896 | 2.95 | 270000 | 1.8213 | 0.6419 |
| 1.8857 | 3.06 | 280000 | 1.8115 | 0.6433 |
| 1.8752 | 3.17 | 290000 | 1.8037 | 0.6443 |
| 1.8662 | 3.28 | 300000 | 1.7949 | 0.6457 |
| 1.8575 | 3.39 | 310000 | 1.7871 | 0.6470 |
| 1.8538 | 3.5 | 320000 | 1.7793 | 0.6478 |
| 1.8426 | 3.6 | 330000 | 1.7734 | 0.6489 |
| 1.8389 | 3.71 | 340000 | 1.7646 | 0.6501 |
| 1.8278 | 3.82 | 350000 | 1.7598 | 0.6511 |
| 1.8319 | 3.93 | 360000 | 1.7529 | 0.6520 |
| 1.8203 | 4.04 | 370000 | 1.7471 | 0.6527 |
| 1.8162 | 4.15 | 380000 | 1.7412 | 0.6536 |
| 1.8113 | 4.26 | 390000 | 1.7373 | 0.6543 |
| 1.8055 | 4.37 | 400000 | 1.7324 | 0.6551 |
| 1.7991 | 4.48 | 410000 | 1.7285 | 0.6556 |
| 1.7965 | 4.59 | 420000 | 1.7246 | 0.6562 |
| 1.7938 | 4.7 | 430000 | 1.7207 | 0.6567 |
| 1.793 | 4.81 | 440000 | 1.7178 | 0.6571 |
| 1.7848 | 4.92 | 450000 | 1.7148 | 0.6576 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.14.0a0+410ce96
- Datasets 2.10.1
- Tokenizers 0.13.2
- Task
- Fill-Mask
- Library
- transformers
- Type
- hybridbert
- Created
- Feb 28, 2023
- Updated
- Mar 6, 2023