update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,95 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- wikitext
|
| 7 |
+
metrics:
|
| 8 |
+
- accuracy
|
| 9 |
+
model-index:
|
| 10 |
+
- name: wikitext_roberta-base
|
| 11 |
+
results:
|
| 12 |
+
- task:
|
| 13 |
+
name: Masked Language Modeling
|
| 14 |
+
type: fill-mask
|
| 15 |
+
dataset:
|
| 16 |
+
name: wikitext
|
| 17 |
+
type: wikitext
|
| 18 |
+
args: wikitext-2-raw-v1
|
| 19 |
+
metrics:
|
| 20 |
+
- name: Accuracy
|
| 21 |
+
type: accuracy
|
| 22 |
+
value: 0.7311184760057123
|
| 23 |
+
---
|
| 24 |
+
|
| 25 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 26 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 27 |
+
|
| 28 |
+
# wikitext_roberta-base
|
| 29 |
+
|
| 30 |
+
This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on the wikitext dataset.
|
| 31 |
+
It achieves the following results on the evaluation set:
|
| 32 |
+
- Loss: 1.2506
|
| 33 |
+
- Accuracy: 0.7311
|
| 34 |
+
|
| 35 |
+
## Model description
|
| 36 |
+
|
| 37 |
+
More information needed
|
| 38 |
+
|
| 39 |
+
## Intended uses & limitations
|
| 40 |
+
|
| 41 |
+
More information needed
|
| 42 |
+
|
| 43 |
+
## Training and evaluation data
|
| 44 |
+
|
| 45 |
+
More information needed
|
| 46 |
+
|
| 47 |
+
## Training procedure
|
| 48 |
+
|
| 49 |
+
### Training hyperparameters
|
| 50 |
+
|
| 51 |
+
The following hyperparameters were used during training:
|
| 52 |
+
- learning_rate: 5e-05
|
| 53 |
+
- train_batch_size: 8
|
| 54 |
+
- eval_batch_size: 8
|
| 55 |
+
- seed: 42
|
| 56 |
+
- gradient_accumulation_steps: 16
|
| 57 |
+
- total_train_batch_size: 128
|
| 58 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 59 |
+
- lr_scheduler_type: linear
|
| 60 |
+
- lr_scheduler_warmup_steps: 50
|
| 61 |
+
- num_epochs: 20.0
|
| 62 |
+
- mixed_precision_training: Native AMP
|
| 63 |
+
|
| 64 |
+
### Training results
|
| 65 |
+
|
| 66 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 67 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 68 |
+
| 1.4175 | 0.99 | 37 | 1.3355 | 0.7194 |
|
| 69 |
+
| 1.438 | 1.99 | 74 | 1.2953 | 0.7249 |
|
| 70 |
+
| 1.4363 | 2.99 | 111 | 1.2759 | 0.7276 |
|
| 71 |
+
| 1.3391 | 3.99 | 148 | 1.2904 | 0.7252 |
|
| 72 |
+
| 1.3741 | 4.99 | 185 | 1.2621 | 0.7290 |
|
| 73 |
+
| 1.2771 | 5.99 | 222 | 1.2312 | 0.7353 |
|
| 74 |
+
| 1.287 | 6.99 | 259 | 1.2542 | 0.7289 |
|
| 75 |
+
| 1.29 | 7.99 | 296 | 1.2290 | 0.7345 |
|
| 76 |
+
| 1.2948 | 8.99 | 333 | 1.2537 | 0.7286 |
|
| 77 |
+
| 1.2741 | 9.99 | 370 | 1.2199 | 0.7354 |
|
| 78 |
+
| 1.2342 | 10.99 | 407 | 1.2520 | 0.7309 |
|
| 79 |
+
| 1.2199 | 11.99 | 444 | 1.2738 | 0.7260 |
|
| 80 |
+
| 1.206 | 12.99 | 481 | 1.2286 | 0.7335 |
|
| 81 |
+
| 1.221 | 13.99 | 518 | 1.2421 | 0.7327 |
|
| 82 |
+
| 1.2062 | 14.99 | 555 | 1.2402 | 0.7328 |
|
| 83 |
+
| 1.2305 | 15.99 | 592 | 1.2473 | 0.7308 |
|
| 84 |
+
| 1.2426 | 16.99 | 629 | 1.2250 | 0.7318 |
|
| 85 |
+
| 1.2096 | 17.99 | 666 | 1.2186 | 0.7353 |
|
| 86 |
+
| 1.1961 | 18.99 | 703 | 1.2214 | 0.7361 |
|
| 87 |
+
| 1.2136 | 19.99 | 740 | 1.2506 | 0.7311 |
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
### Framework versions
|
| 91 |
+
|
| 92 |
+
- Transformers 4.21.0.dev0
|
| 93 |
+
- Pytorch 1.11.0+cu113
|
| 94 |
+
- Datasets 2.3.3.dev0
|
| 95 |
+
- Tokenizers 0.12.1
|