Commit
•
309fdc3
1
Parent(s):
34beb24
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: moses_cbgpt
|
| 7 |
+
results: []
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 11 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 12 |
+
|
| 13 |
+
# moses_cbgpt
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on an unknown dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.5234
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 0.0004
|
| 37 |
+
- train_batch_size: 32
|
| 38 |
+
- eval_batch_size: 32
|
| 39 |
+
- seed: 42
|
| 40 |
+
- gradient_accumulation_steps: 8
|
| 41 |
+
- total_train_batch_size: 256
|
| 42 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
+
- lr_scheduler_type: cosine
|
| 44 |
+
- lr_scheduler_warmup_steps: 1000
|
| 45 |
+
- num_epochs: 2
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
| 50 |
+
|:-------------:|:-----:|:-----:|:---------------:|
|
| 51 |
+
| 0.9833 | 0.18 | 1000 | 0.7188 |
|
| 52 |
+
| 0.6852 | 0.36 | 2000 | 0.6362 |
|
| 53 |
+
| 0.6347 | 0.54 | 3000 | 0.6076 |
|
| 54 |
+
| 0.6094 | 0.72 | 4000 | 0.5887 |
|
| 55 |
+
| 0.5908 | 0.9 | 5000 | 0.5725 |
|
| 56 |
+
| 0.5743 | 1.08 | 6000 | 0.5608 |
|
| 57 |
+
| 0.5606 | 1.26 | 7000 | 0.5483 |
|
| 58 |
+
| 0.5492 | 1.44 | 8000 | 0.5373 |
|
| 59 |
+
| 0.5396 | 1.62 | 9000 | 0.5298 |
|
| 60 |
+
| 0.5327 | 1.79 | 10000 | 0.5248 |
|
| 61 |
+
| 0.5289 | 1.97 | 11000 | 0.5234 |
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
### Framework versions
|
| 65 |
+
|
| 66 |
+
- Transformers 4.30.2
|
| 67 |
+
- Pytorch 2.0.0
|
| 68 |
+
- Datasets 2.1.0
|
| 69 |
+
- Tokenizers 0.13.3
|