holylovenia
commited on
Commit
•
c3e9d45
1
Parent(s):
4d9ec89
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -13,8 +13,8 @@ tags:
|
|
13 |
- image-captioning
|
14 |
---
|
15 |
|
16 |
-
|
17 |
-
|
18 |
|
19 |
|
20 |
This is a local dataset. You have to obtain this dataset separately from [{homepage}]({homepage}) to use this dataloader.
|
@@ -26,25 +26,25 @@ fil, ind, tha, vie
|
|
26 |
## Supported Tasks
|
27 |
|
28 |
Image Captioning
|
29 |
-
|
30 |
## Dataset Usage
|
31 |
### Using `datasets` library
|
32 |
```
|
33 |
-
|
34 |
-
|
35 |
```
|
36 |
### Using `seacrowd` library
|
37 |
```import seacrowd as sc
|
38 |
# Load the dataset using the default config
|
39 |
-
|
40 |
# Check all available subsets (config names) of the dataset
|
41 |
-
|
42 |
# Load the dataset using a specific config
|
43 |
-
|
44 |
```
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
|
49 |
## Dataset Homepage
|
50 |
|
|
|
13 |
- image-captioning
|
14 |
---
|
15 |
|
16 |
+
CC3M-35L is created by translating Conceptual Captions 3M (Sharma et al., 2018),
|
17 |
+
originally in English, to the other 34 languages using Google's machine translation API.
|
18 |
|
19 |
|
20 |
This is a local dataset. You have to obtain this dataset separately from [{homepage}]({homepage}) to use this dataloader.
|
|
|
26 |
## Supported Tasks
|
27 |
|
28 |
Image Captioning
|
29 |
+
|
30 |
## Dataset Usage
|
31 |
### Using `datasets` library
|
32 |
```
|
33 |
+
from datasets import load_dataset
|
34 |
+
dset = datasets.load_dataset("SEACrowd/cc3m_35l", trust_remote_code=True)
|
35 |
```
|
36 |
### Using `seacrowd` library
|
37 |
```import seacrowd as sc
|
38 |
# Load the dataset using the default config
|
39 |
+
dset = sc.load_dataset("cc3m_35l", schema="seacrowd")
|
40 |
# Check all available subsets (config names) of the dataset
|
41 |
+
print(sc.available_config_names("cc3m_35l"))
|
42 |
# Load the dataset using a specific config
|
43 |
+
dset = sc.load_dataset_by_config_name(config_name="<config_name>")
|
44 |
```
|
45 |
+
|
46 |
+
More details on how to load the `seacrowd` library can be found [here](https://github.com/SEACrowd/seacrowd-datahub?tab=readme-ov-file#how-to-use).
|
47 |
+
|
48 |
|
49 |
## Dataset Homepage
|
50 |
|