|
--- |
|
tags: |
|
- text2text-generation |
|
- definition-modeling |
|
metrics: |
|
- rouge |
|
model-index: |
|
- name: mt0-definition-ru-xl |
|
results: [] |
|
language: |
|
- ru |
|
widget: |
|
- text: "Мы сели в тачку и поехали по ресторанам. Что такое тачка?" |
|
example_title: "Definition generation" |
|
--- |
|
|
|
# mt0-definition-ru-xl |
|
|
|
This model is a version of [mt0-xl](https://huggingface.co/bigscience/mt0-xl) finetuned on the Russian part of CoDWoE dataset. |
|
|
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.6241 |
|
- Rouge1: 0.2536 |
|
- Rouge2: 0.003 |
|
- Rougel: 0.2531 |
|
- Rougelsum: 0.2527 |
|
- Gen Len: 24.0693 |
|
|
|
## 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: 4 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 8 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 128 |
|
- total_eval_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 20.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:| |
|
| 2.0449 | 1.0 | 512 | 1.6755 | 0.0817 | 0.0 | 0.0778 | 0.0817 | 16.7581 | |
|
| 1.707 | 2.0 | 1025 | 1.6182 | 0.096 | 0.0 | 0.097 | 0.1 | 15.8621 | |
|
| 1.5398 | 3.0 | 1537 | 1.6085 | 0.1394 | 0.0034 | 0.1401 | 0.1416 | 16.4765 | |
|
| 1.4142 | 4.0 | 2050 | 1.6016 | 0.1132 | 0.0 | 0.1132 | 0.1098 | 16.2732 | |
|
| 1.3102 | 5.0 | 2562 | 1.6241 | 0.2082 | 0.0034 | 0.2054 | 0.2061 | 16.2877 | |
|
| 1.2162 | 6.0 | 3075 | 1.6281 | 0.1549 | 0.0 | 0.1549 | 0.1549 | 16.1581 | |
|
| 1.1364 | 7.0 | 3587 | 1.6622 | 0.1583 | 0.0 | 0.1575 | 0.1589 | 15.9925 | |
|
| 1.0649 | 8.0 | 4100 | 1.6812 | 0.2033 | 0.0137 | 0.2012 | 0.2027 | 16.5099 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.30.2 |
|
- Pytorch 1.13.1+rocm5.2 |
|
- Datasets 2.12.0 |
|
- Tokenizers 0.12.1 |
|
|
