oMateos2020
commited on
Commit
•
e12758d
1
Parent(s):
cdddec8
update model card README.md
Browse files
README.md
CHANGED
@@ -18,7 +18,7 @@ model-index:
|
|
18 |
metrics:
|
19 |
- name: Rouge1
|
20 |
type: rouge
|
21 |
-
value:
|
22 |
---
|
23 |
|
24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -26,14 +26,14 @@ should probably proofread and complete it, then remove this comment. -->
|
|
26 |
|
27 |
# pegasus-newsroom-cnn_full-adafactor-bs6
|
28 |
|
29 |
-
This model
|
30 |
It achieves the following results on the evaluation set:
|
31 |
-
- Loss:
|
32 |
-
- Rouge1:
|
33 |
-
- Rouge2:
|
34 |
-
- Rougel:
|
35 |
-
- Rougelsum:
|
36 |
-
- Gen Len:
|
37 |
|
38 |
## Model description
|
39 |
|
@@ -52,15 +52,15 @@ More information needed
|
|
52 |
### Training hyperparameters
|
53 |
|
54 |
The following hyperparameters were used during training:
|
55 |
-
- learning_rate:
|
56 |
-
- train_batch_size:
|
57 |
-
- eval_batch_size:
|
58 |
- seed: 42
|
59 |
-
- gradient_accumulation_steps:
|
60 |
-
- total_train_batch_size:
|
61 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
62 |
- lr_scheduler_type: linear
|
63 |
-
- lr_scheduler_warmup_steps:
|
64 |
- num_epochs: 1
|
65 |
- mixed_precision_training: Native AMP
|
66 |
- label_smoothing_factor: 0.1
|
@@ -69,11 +69,8 @@ The following hyperparameters were used during training:
|
|
69 |
|
70 |
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|
71 |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
|
72 |
-
|
|
73 |
-
|
|
74 |
-
| 3.1017 | 0.3 | 897 | 2.9891 | 39.0977 | 17.9198 | 27.9078 | 36.2363 | 58.5172 |
|
75 |
-
| 3.2891 | 0.4 | 1196 | 3.5756 | 29.5555 | 11.7552 | 22.4675 | 27.2432 | 45.0232 |
|
76 |
-
| 637.0317 | 0.5 | 1495 | nan | 0.0 | 0.0 | 0.0 | 0.0 | 1.0 |
|
77 |
|
78 |
|
79 |
### Framework versions
|
|
|
18 |
metrics:
|
19 |
- name: Rouge1
|
20 |
type: rouge
|
21 |
+
value: 44.1026
|
22 |
---
|
23 |
|
24 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
26 |
|
27 |
# pegasus-newsroom-cnn_full-adafactor-bs6
|
28 |
|
29 |
+
This model is a fine-tuned version of [oMateos2020/pegasus-newsroom-cnn_full-adafactor-bs6](https://huggingface.co/oMateos2020/pegasus-newsroom-cnn_full-adafactor-bs6) on the cnn_dailymail dataset.
|
30 |
It achieves the following results on the evaluation set:
|
31 |
+
- Loss: 2.8671
|
32 |
+
- Rouge1: 44.1026
|
33 |
+
- Rouge2: 21.4261
|
34 |
+
- Rougel: 31.2033
|
35 |
+
- Rougelsum: 41.0324
|
36 |
+
- Gen Len: 72.0839
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
52 |
### Training hyperparameters
|
53 |
|
54 |
The following hyperparameters were used during training:
|
55 |
+
- learning_rate: 6.4e-05
|
56 |
+
- train_batch_size: 4
|
57 |
+
- eval_batch_size: 4
|
58 |
- seed: 42
|
59 |
+
- gradient_accumulation_steps: 64
|
60 |
+
- total_train_batch_size: 256
|
61 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
62 |
- lr_scheduler_type: linear
|
63 |
+
- lr_scheduler_warmup_steps: 500
|
64 |
- num_epochs: 1
|
65 |
- mixed_precision_training: Native AMP
|
66 |
- label_smoothing_factor: 0.1
|
|
|
69 |
|
70 |
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|
71 |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:-------:|
|
72 |
+
| 2.9343 | 0.5 | 560 | 2.8733 | 44.1226 | 21.4087 | 31.2431 | 41.0683 | 69.367 |
|
73 |
+
| 2.9855 | 1.0 | 1120 | 2.8671 | 44.1026 | 21.4261 | 31.2033 | 41.0324 | 72.0839 |
|
|
|
|
|
|
|
74 |
|
75 |
|
76 |
### Framework versions
|