Training complete
Browse files- README.md +2 -17
- generation_config.json +1 -0
README.md
CHANGED
|
@@ -4,8 +4,6 @@ base_model: google-t5/t5-small
|
|
| 4 |
tags:
|
| 5 |
- summarization
|
| 6 |
- generated_from_trainer
|
| 7 |
-
metrics:
|
| 8 |
-
- rouge
|
| 9 |
model-index:
|
| 10 |
- name: xsum_unaligned_smallT5
|
| 11 |
results: []
|
|
@@ -17,12 +15,6 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
# xsum_unaligned_smallT5
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [google-t5/t5-small](https://huggingface.co/google-t5/t5-small) on an unknown dataset.
|
| 20 |
-
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 2.8386
|
| 22 |
-
- Rouge1: 0.2219
|
| 23 |
-
- Rouge2: 0.0465
|
| 24 |
-
- Rougel: 0.1675
|
| 25 |
-
- Rougelsum: 0.1714
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
|
@@ -49,18 +41,11 @@ The following hyperparameters were used during training:
|
|
| 49 |
- total_train_batch_size: 16
|
| 50 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 51 |
- lr_scheduler_type: linear
|
| 52 |
-
- training_steps:
|
|
|
|
| 53 |
|
| 54 |
### Training results
|
| 55 |
|
| 56 |
-
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|
| 57 |
-
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|
|
| 58 |
-
| No log | 0.86 | 3 | 3.3139 | 0.2048 | 0.0374 | 0.1476 | 0.1663 |
|
| 59 |
-
| 3.6766 | 2.0 | 7 | 3.0257 | 0.2078 | 0.0347 | 0.1515 | 0.1644 |
|
| 60 |
-
| 3.6766 | 2.86 | 10 | 2.9150 | 0.2208 | 0.0427 | 0.1660 | 0.1761 |
|
| 61 |
-
| 3.0126 | 4.0 | 14 | 2.8600 | 0.2229 | 0.0449 | 0.1704 | 0.1782 |
|
| 62 |
-
| 3.0126 | 4.86 | 17 | 2.8455 | 0.2184 | 0.0455 | 0.1649 | 0.1689 |
|
| 63 |
-
| 2.8062 | 5.71 | 20 | 2.8386 | 0.2219 | 0.0465 | 0.1675 | 0.1714 |
|
| 64 |
|
| 65 |
|
| 66 |
### Framework versions
|
|
|
|
| 4 |
tags:
|
| 5 |
- summarization
|
| 6 |
- generated_from_trainer
|
|
|
|
|
|
|
| 7 |
model-index:
|
| 8 |
- name: xsum_unaligned_smallT5
|
| 9 |
results: []
|
|
|
|
| 15 |
# xsum_unaligned_smallT5
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [google-t5/t5-small](https://huggingface.co/google-t5/t5-small) on an unknown dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
|
|
| 41 |
- total_train_batch_size: 16
|
| 42 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
- lr_scheduler_type: linear
|
| 44 |
+
- training_steps: 200000
|
| 45 |
+
- mixed_precision_training: Native AMP
|
| 46 |
|
| 47 |
### Training results
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
|
| 50 |
|
| 51 |
### Framework versions
|
generation_config.json
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"decoder_start_token_id": 0,
|
| 3 |
"eos_token_id": 1,
|
| 4 |
"pad_token_id": 0,
|
|
|
|
| 1 |
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
"decoder_start_token_id": 0,
|
| 4 |
"eos_token_id": 1,
|
| 5 |
"pad_token_id": 0,
|