File size: 2,055 Bytes
e0f65e0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 |
---
license: apache-2.0
base_model: google/mt5-large
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: MT5_large_NO_CNN-idun-final
results: []
---
<!-- 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. -->
# MT5_large_NO_CNN-idun-final
This model is a fine-tuned version of [google/mt5-large](https://huggingface.co/google/mt5-large) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.8492
- Rouge1: 31.9047
- Rouge2: 12.0487
- Rougel: 21.7323
- Rougelsum: 29.4557
- Gen Len: 98.7777
## 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: 2
- eval_batch_size: 2
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
| 2.246 | 1.0 | 6118 | 1.9772 | 30.9142 | 11.28 | 21.0914 | 28.4499 | 95.2396 |
| 2.0899 | 2.0 | 12236 | 1.9070 | 31.3973 | 11.6219 | 21.3122 | 28.9365 | 100.0907 |
| 1.9736 | 3.0 | 18354 | 1.8716 | 31.5752 | 11.7955 | 21.4748 | 29.1354 | 100.4973 |
| 1.9189 | 4.0 | 24472 | 1.8547 | 31.9802 | 12.183 | 21.8505 | 29.5171 | 98.4764 |
| 1.8697 | 5.0 | 30590 | 1.8492 | 31.9047 | 12.0487 | 21.7323 | 29.4557 | 98.7777 |
### Framework versions
- Transformers 4.32.1
- Pytorch 2.3.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
|