Update README.md
Browse files
README.md
CHANGED
@@ -1,54 +1,62 @@
|
|
1 |
---
|
2 |
language:
|
3 |
- rus
|
|
|
4 |
tags:
|
5 |
- mbart
|
6 |
inference:
|
7 |
parameters:
|
8 |
no_repeat_ngram_size: 4,
|
9 |
-
num_beams
|
10 |
datasets:
|
11 |
- IlyaGusev/gazeta
|
12 |
- samsum
|
13 |
- samsum_(translated_into_Russian)
|
14 |
widget:
|
15 |
-
- text:
|
16 |
Джефф: Могу ли я обучить модель 🤗 Transformers на Amazon SageMaker?
|
17 |
-
|
|
|
|
|
|
|
18 |
Джефф: Хорошо.
|
|
|
19 |
Джефф: и как я могу начать?
|
|
|
20 |
Джефф: где я могу найти документацию?
|
21 |
-
Филипп: ок, ок, здесь можно найти все: https://huggingface.co/blog/the-partnership-amazon-sagemaker-and-hugging-face
|
22 |
|
|
|
|
|
23 |
model-index:
|
24 |
-
- name:
|
25 |
results:
|
26 |
-
- task:
|
27 |
name: Abstractive Dialogue Summarization
|
28 |
-
type: abstractive-text-summarization
|
29 |
dataset:
|
30 |
-
name:
|
31 |
type: samsum
|
32 |
metrics:
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
|
|
45 |
---
|
46 |
### 📝 Description
|
47 |
|
48 |
MBart for Russian summarization fine-tuned for **dialogues** summarization.
|
49 |
|
50 |
|
51 |
-
This model was firstly fine-tuned by [Ilya Gusev](https://hf.co/IlyaGusev) on [Gazeta dataset](https://huggingface.co/datasets/IlyaGusev/gazeta). We have **fine tuned** that model on [SamSum dataset]() **translated to Russian** using GoogleTranslateAPI
|
52 |
|
53 |
🤗 Moreover! We have implemented a **! telegram bot [@summarization_bot](https://t.me/summarization_bot) !** with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages! 🤗
|
54 |
|
@@ -83,4 +91,4 @@ output_ids = model.generate(
|
|
83 |
|
84 |
summary = tokenizer.decode(output_ids, skip_special_tokens=True)
|
85 |
print(summary)
|
86 |
-
```
|
|
|
1 |
---
|
2 |
language:
|
3 |
- rus
|
4 |
+
- ru
|
5 |
tags:
|
6 |
- mbart
|
7 |
inference:
|
8 |
parameters:
|
9 |
no_repeat_ngram_size: 4,
|
10 |
+
num_beams: 5
|
11 |
datasets:
|
12 |
- IlyaGusev/gazeta
|
13 |
- samsum
|
14 |
- samsum_(translated_into_Russian)
|
15 |
widget:
|
16 |
+
- text: >
|
17 |
Джефф: Могу ли я обучить модель 🤗 Transformers на Amazon SageMaker?
|
18 |
+
|
19 |
+
Филипп: Конечно, вы можете использовать новый контейнер для глубокого
|
20 |
+
обучения HuggingFace.
|
21 |
+
|
22 |
Джефф: Хорошо.
|
23 |
+
|
24 |
Джефф: и как я могу начать?
|
25 |
+
|
26 |
Джефф: где я могу найти документацию?
|
|
|
27 |
|
28 |
+
Филипп: ок, ок, здесь можно найти все:
|
29 |
+
https://huggingface.co/blog/the-partnership-amazon-sagemaker-and-hugging-face
|
30 |
model-index:
|
31 |
+
- name: mbart_ruDialogSum
|
32 |
results:
|
33 |
+
- task:
|
34 |
name: Abstractive Dialogue Summarization
|
35 |
+
type: abstractive-text-summarization
|
36 |
dataset:
|
37 |
+
name: SAMSum Corpus (translated to Russian)
|
38 |
type: samsum
|
39 |
metrics:
|
40 |
+
- name: Validation ROGUE-1
|
41 |
+
type: rogue-1
|
42 |
+
value: 34.5
|
43 |
+
- name: Validation ROGUE-L
|
44 |
+
type: rogue-l
|
45 |
+
value: 33
|
46 |
+
- name: Test ROGUE-1
|
47 |
+
type: rogue-1
|
48 |
+
value: 31
|
49 |
+
- name: Test ROGUE-L
|
50 |
+
type: rogue-l
|
51 |
+
value: 28
|
52 |
+
license: cc
|
53 |
---
|
54 |
### 📝 Description
|
55 |
|
56 |
MBart for Russian summarization fine-tuned for **dialogues** summarization.
|
57 |
|
58 |
|
59 |
+
This model was firstly fine-tuned by [Ilya Gusev](https://hf.co/IlyaGusev) on [Gazeta dataset](https://huggingface.co/datasets/IlyaGusev/gazeta). We have **fine tuned** that model on [SamSum dataset](https://huggingface.co/datasets/samsum) **translated to Russian** using GoogleTranslateAPI
|
60 |
|
61 |
🤗 Moreover! We have implemented a **! telegram bot [@summarization_bot](https://t.me/summarization_bot) !** with the inference of this model. Add it to the chat and get summaries instead of dozens spam messages! 🤗
|
62 |
|
|
|
91 |
|
92 |
summary = tokenizer.decode(output_ids, skip_special_tokens=True)
|
93 |
print(summary)
|
94 |
+
```
|