Commit
•
5b10a21
1
Parent(s):
a61ad2c
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- generated_from_trainer
|
| 4 |
+
model-index:
|
| 5 |
+
- name: legal-french-roberta-base
|
| 6 |
+
results: []
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 10 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 11 |
+
|
| 12 |
+
# legal-french-roberta-base
|
| 13 |
+
|
| 14 |
+
This model was trained from scratch on an unknown dataset.
|
| 15 |
+
It achieves the following results on the evaluation set:
|
| 16 |
+
- Loss: 0.4293
|
| 17 |
+
|
| 18 |
+
## Model description
|
| 19 |
+
|
| 20 |
+
More information needed
|
| 21 |
+
|
| 22 |
+
## Intended uses & limitations
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Training and evaluation data
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training procedure
|
| 31 |
+
|
| 32 |
+
### Training hyperparameters
|
| 33 |
+
|
| 34 |
+
The following hyperparameters were used during training:
|
| 35 |
+
- learning_rate: 0.0001
|
| 36 |
+
- train_batch_size: 16
|
| 37 |
+
- eval_batch_size: 16
|
| 38 |
+
- seed: 42
|
| 39 |
+
- distributed_type: tpu
|
| 40 |
+
- num_devices: 8
|
| 41 |
+
- gradient_accumulation_steps: 4
|
| 42 |
+
- total_train_batch_size: 512
|
| 43 |
+
- total_eval_batch_size: 128
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: cosine
|
| 46 |
+
- lr_scheduler_warmup_ratio: 0.05
|
| 47 |
+
- training_steps: 1000000
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 52 |
+
|:-------------:|:-----:|:-------:|:---------------:|
|
| 53 |
+
| 0.8649 | 0.05 | 50000 | 0.7819 |
|
| 54 |
+
| 0.7852 | 0.1 | 100000 | 0.6027 |
|
| 55 |
+
| 0.5898 | 1.02 | 150000 | 0.5842 |
|
| 56 |
+
| 0.6136 | 1.07 | 200000 | 0.5343 |
|
| 57 |
+
| 0.6135 | 1.12 | 250000 | 0.5461 |
|
| 58 |
+
| 0.5804 | 2.03 | 300000 | 0.5295 |
|
| 59 |
+
| 0.5602 | 2.08 | 350000 | 0.5120 |
|
| 60 |
+
| 0.5446 | 2.13 | 400000 | 0.4904 |
|
| 61 |
+
| 0.5414 | 3.05 | 450000 | 0.4853 |
|
| 62 |
+
| 0.5765 | 3.1 | 500000 | 0.4788 |
|
| 63 |
+
| 0.6903 | 4.01 | 550000 | 0.4597 |
|
| 64 |
+
| 0.6149 | 4.06 | 600000 | 0.4556 |
|
| 65 |
+
| 0.5649 | 4.11 | 650000 | 0.4543 |
|
| 66 |
+
| 0.6449 | 5.03 | 700000 | 0.4489 |
|
| 67 |
+
| 0.6425 | 5.08 | 750000 | 0.4386 |
|
| 68 |
+
| 0.6263 | 5.13 | 800000 | 0.4344 |
|
| 69 |
+
| 0.6035 | 6.05 | 850000 | 0.4317 |
|
| 70 |
+
| 0.607 | 6.1 | 900000 | 0.4332 |
|
| 71 |
+
| 0.5899 | 7.01 | 950000 | 0.4321 |
|
| 72 |
+
| 0.5751 | 7.06 | 1000000 | 0.4293 |
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
### Framework versions
|
| 76 |
+
|
| 77 |
+
- Transformers 4.20.1
|
| 78 |
+
- Pytorch 1.12.0+cu102
|
| 79 |
+
- Datasets 2.8.0
|
| 80 |
+
- Tokenizers 0.12.1
|