Update README.md
Browse files
README.md
CHANGED
@@ -144,9 +144,9 @@ configs:
|
|
144 |
path: demographics/train-*
|
145 |
tags: []
|
146 |
---
|
147 |
-
 from Cohere
|
150 |
This dataset can be used to train, finetune, and evaluate multilingual LLMs.
|
151 |
|
152 |
- **Curated by:** Contributors of [Aya Open Science Intiative](https://aya.for.ai/).
|
@@ -156,11 +156,11 @@ This dataset can be used to train, finetune, and evaluate multilingual LLMs.
|
|
156 |
|
157 |
| Name | Explanation |
|
158 |
|------|--------------|
|
159 |
-
| [aya_dataset](https://huggingface.co/datasets/
|
160 |
-
| [aya_collection](https://huggingface.co/datasets/
|
161 |
-
| [aya_collection_language_split](https://huggingface.co/datasets/
|
162 |
-
| [aya_evaluation_suite](https://huggingface.co/datasets/
|
163 |
-
| [aya_redteaming](https://huggingface.co/datasets/
|
164 |
|
165 |
# Dataset
|
166 |
The `Aya Dataset` comprises of two types of data:
|
@@ -174,10 +174,10 @@ To load this dataset consisting of both prompt-completions and demographics data
|
|
174 |
from datasets import load_dataset
|
175 |
|
176 |
# Load the annotations dataset
|
177 |
-
aya_dataset = load_dataset("
|
178 |
|
179 |
# Load the demographics dataset
|
180 |
-
aya_demographics = load_dataset("
|
181 |
```
|
182 |
|
183 |
## Data Fields
|
@@ -359,7 +359,7 @@ The dataset covers 65 languages: 28 high-resource, 12 mid-resource, and 31 low-r
|
|
359 |
|
360 |
|
361 |
## Authorship
|
362 |
-
- **Publishing Organization:** [Cohere
|
363 |
- **Industry Type:** Not-for-profit - Tech
|
364 |
- **Contact Details:** https://aya.for.ai/
|
365 |
|
|
|
144 |
path: demographics/train-*
|
145 |
tags: []
|
146 |
---
|
147 |
+

|
148 |
# Dataset Summary
|
149 |
+
The `Aya Dataset` is a multilingual instruction fine-tuning dataset curated by an open-science community via [Aya Annotation Platform](https://aya.for.ai/) from Cohere Labs. The dataset contains a total of 204k human-annotated prompt-completion pairs along with the demographics data of the annotators.<br>
|
150 |
This dataset can be used to train, finetune, and evaluate multilingual LLMs.
|
151 |
|
152 |
- **Curated by:** Contributors of [Aya Open Science Intiative](https://aya.for.ai/).
|
|
|
156 |
|
157 |
| Name | Explanation |
|
158 |
|------|--------------|
|
159 |
+
| [aya_dataset](https://huggingface.co/datasets/CohereLabs/aya_dataset) | Human-annotated multilingual instruction finetuning dataset, comprising over 204K instances across 65 languages. |
|
160 |
+
| [aya_collection](https://huggingface.co/datasets/CohereLabs/aya_collection) | Created by applying instruction-style templates from fluent speakers to 44 datasets, including translations of 19 instruction-style datasets into 101 languages, providing 513M instances for various tasks.|
|
161 |
+
| [aya_collection_language_split](https://huggingface.co/datasets/CohereLabs/aya_collection_language_split) | Aya Collection structured based on language level subsets. |
|
162 |
+
| [aya_evaluation_suite](https://huggingface.co/datasets/CohereLabs/aya_evaluation_suite) | A diverse evaluation set for multilingual open-ended generation, featuring 250 culturally grounded prompts in 7 languages, 200 translated prompts in 24 languages, and human-edited versions selected for cross-cultural relevance from English Dolly in 6 languages.|
|
163 |
+
| [aya_redteaming](https://huggingface.co/datasets/CohereLabs/aya_redteaming)| A red-teaming dataset consisting of harmful prompts in 8 languages across 9 different categories of harm with explicit labels for "global" and "local" harm.|
|
164 |
|
165 |
# Dataset
|
166 |
The `Aya Dataset` comprises of two types of data:
|
|
|
174 |
from datasets import load_dataset
|
175 |
|
176 |
# Load the annotations dataset
|
177 |
+
aya_dataset = load_dataset("CohereLabs/aya_dataset")
|
178 |
|
179 |
# Load the demographics dataset
|
180 |
+
aya_demographics = load_dataset("CohereLabs/aya_dataset", "demographics")
|
181 |
```
|
182 |
|
183 |
## Data Fields
|
|
|
359 |
|
360 |
|
361 |
## Authorship
|
362 |
+
- **Publishing Organization:** [Cohere Labs](https://cohere.com/research)
|
363 |
- **Industry Type:** Not-for-profit - Tech
|
364 |
- **Contact Details:** https://aya.for.ai/
|
365 |
|