Update README.md
Browse files
README.md
CHANGED
@@ -26,34 +26,65 @@ model-index:
|
|
26 |
|
27 |
---
|
28 |
|
29 |
-
This markdown file contains the spec for the modelcard metadata regarding evaluation parameters. When present, and only then, 'model-index', 'datasets' and 'license' contents will be verified when git pushing changes to your README.md file.
|
30 |
-
Valid license identifiers can be found in [our docs](https://huggingface.co/docs/hub/repositories-licenses).
|
31 |
-
|
32 |
-
For the full model card template, see: [modelcard_template.md file](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/modelcard_template.md).
|
33 |
-
---
|
34 |
-
|
35 |
-
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
36 |
-
should probably proofread and complete it, then remove this comment. -->
|
37 |
-
|
38 |
# longformer_4096_qsi
|
39 |
|
40 |
-
This model is a fine-tuned version of [mrm8488/longformer-base-4096-finetuned-squadv2](https://huggingface.co/mrm8488/longformer-base-4096-finetuned-squadv2) on
|
41 |
It achieves the following results on the evaluation set:
|
42 |
- Loss: 2.9598
|
43 |
|
44 |
## Model description
|
45 |
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
51 |
|
52 |
## Training and evaluation data
|
53 |
|
54 |
-
|
|
|
|
|
55 |
|
56 |
-
|
57 |
|
58 |
### Training hyperparameters
|
59 |
|
|
|
26 |
|
27 |
---
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
# longformer_4096_qsi
|
30 |
|
31 |
+
This model is a fine-tuned version of [mrm8488/longformer-base-4096-finetuned-squadv2](https://huggingface.co/mrm8488/longformer-base-4096-finetuned-squadv2) on a tiny [NovelQSI](https://huggingface.co/datasets/Kkordik/NovelQSI) dataset.
|
32 |
It achieves the following results on the evaluation set:
|
33 |
- Loss: 2.9598
|
34 |
|
35 |
## Model description
|
36 |
|
37 |
+
This model is test model for my research project. The idea of the model is to understand which novel character said the requested quote.
|
38 |
+
It achieves a bit better results on the ´test´ split of the NovelQSI dataset than base longformer-base-4096-finetuned-squadv2 model on the same dataset split.
|
39 |
+
|
40 |
+
**Base model results:**
|
41 |
+
|
42 |
+
```
|
43 |
+
{
|
44 |
+
"exact_match": {
|
45 |
+
"confidence_interval": [8.754452551305853, 14.718614718614718],
|
46 |
+
"score": 12.121212121212121,
|
47 |
+
"standard_error": 1.8579217243778676
|
48 |
+
},
|
49 |
+
"f1": {
|
50 |
+
"confidence_interval": [18.469101076147584, 28.28409063313956],
|
51 |
+
"score": 22.799422799422796,
|
52 |
+
"standard_error": 2.896728175757627
|
53 |
+
},
|
54 |
+
"latency_in_seconds": 0.7730605573419919,
|
55 |
+
"samples_per_second": 1.2935597224598967,
|
56 |
+
"total_time_in_seconds": 178.5769887460001
|
57 |
+
}
|
58 |
+
|
59 |
+
```
|
60 |
+
|
61 |
+
**Achieved results:**
|
62 |
+
|
63 |
+
```
|
64 |
+
{
|
65 |
+
"exact_match": {
|
66 |
+
"confidence_interval": [16.017316017316016, 24.242424242424242],
|
67 |
+
"score": 20.346320346320347,
|
68 |
+
"standard_error": 2.9434375492784994
|
69 |
+
},
|
70 |
+
"f1": {
|
71 |
+
"confidence_interval": [23.123469058324783, 31.823648733317036],
|
72 |
+
"score": 26.580086580086572,
|
73 |
+
"standard_error": 2.593030474995015
|
74 |
+
},
|
75 |
+
"latency_in_seconds": 0.8093855569913422,
|
76 |
+
"samples_per_second": 1.235505120349827,
|
77 |
+
"total_time_in_seconds": 186.96806366500005
|
78 |
+
}
|
79 |
+
```
|
80 |
|
81 |
## Training and evaluation data
|
82 |
|
83 |
+
You can find training code in the github repo of my research:
|
84 |
+
|
85 |
+
https://github.com/Kkordik/NovelQSI/tree/main
|
86 |
|
87 |
+
It was trained and evaluated in notebooks, so it is easy to reproduce.
|
88 |
|
89 |
### Training hyperparameters
|
90 |
|