Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
dataset_name: PashtoNanoChat
|
| 3 |
+
language:
|
| 4 |
+
- ps
|
| 5 |
+
tags:
|
| 6 |
+
- pashto
|
| 7 |
+
- dialogue
|
| 8 |
+
- chatbot
|
| 9 |
+
- nlp
|
| 10 |
+
license: apache-2.0
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# PashtoNanoChat
|
| 14 |
+
|
| 15 |
+
High‑quality Pashto conversational dataset designed for chatbot training, dialogue modeling, and Pashto LLM alignment.
|
| 16 |
+
This dataset contains structured Pashto chat pairs in OpenAI/Ministral message format.
|
| 17 |
+
|
| 18 |
+
## Dataset Summary
|
| 19 |
+
|
| 20 |
+
`PashtoNanoChat` provides clean Pashto dialogues formatted as:
|
| 21 |
+
|
| 22 |
+
```json
|
| 23 |
+
{
|
| 24 |
+
"messages": [
|
| 25 |
+
{"role": "user", "content": "..."},
|
| 26 |
+
{"role": "assistant", "content": "..."}
|
| 27 |
+
]
|
| 28 |
+
}
|
| 29 |
+
```
|
| 30 |
+
|
| 31 |
+
The dataset is suitable for:
|
| 32 |
+
|
| 33 |
+
- Pashto chatbots
|
| 34 |
+
- Instruction‑tuned LLMs
|
| 35 |
+
- Dialogue agents
|
| 36 |
+
- Cultural alignment
|
| 37 |
+
- Low‑resource NLP research
|
| 38 |
+
|
| 39 |
+
## Data Format
|
| 40 |
+
|
| 41 |
+
Each entry contains:
|
| 42 |
+
|
| 43 |
+
| Field | Description |
|
| 44 |
+
|----------|--------------------------|
|
| 45 |
+
| messages | List of chat turns |
|
| 46 |
+
| role | `"user"` or `"assistant"`|
|
| 47 |
+
| content | Pashto text |
|
| 48 |
+
|
| 49 |
+
Example:
|
| 50 |
+
|
| 51 |
+
```json
|
| 52 |
+
{
|
| 53 |
+
"messages": [
|
| 54 |
+
{"role": "user", "content": "سلام، څنګه یې؟"},
|
| 55 |
+
{"role": "assistant", "content": "زه ښه یم، مننه! څنګه مرسته درسره وکړم؟"}
|
| 56 |
+
]
|
| 57 |
+
}
|
| 58 |
+
```
|
| 59 |
+
|
| 60 |
+
## Usage
|
| 61 |
+
|
| 62 |
+
### Load with `datasets`
|
| 63 |
+
|
| 64 |
+
```python
|
| 65 |
+
from datasets import load_dataset
|
| 66 |
+
ds = load_dataset("nassimjp/PashtoNanoChat")
|
| 67 |
+
```
|
| 68 |
+
|
| 69 |
+
### Training Example (Ministral / OpenAI format)
|
| 70 |
+
|
| 71 |
+
```python
|
| 72 |
+
tokenizer.apply_chat_template(
|
| 73 |
+
example["messages"],
|
| 74 |
+
tokenize=True,
|
| 75 |
+
add_generation_prompt=False
|
| 76 |
+
)
|
| 77 |
+
```
|
| 78 |
+
|
| 79 |
+
## License — Apache License 2.0
|
| 80 |
+
|
| 81 |
+
Copyright 2026 NASIMJP
|
| 82 |
+
|
| 83 |
+
Licensed under the Apache License, Version 2.0 (the "License");
|
| 84 |
+
you may not use this dataset except in compliance with the License.
|
| 85 |
+
You may obtain a copy of the License at:
|
| 86 |
+
|
| 87 |
+
`http://www.apache.org/licenses/LICENSE-2.0` [(apache.org in Bing)](https://www.bing.com/search?q="http%3A%2F%2Fwww.apache.org%2Flicenses%2FLICENSE-2.0")
|
| 88 |
+
|
| 89 |
+
Unless required by applicable law or agreed to in writing, data
|
| 90 |
+
distributed under the License is distributed on an "AS IS" BASIS,
|
| 91 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 92 |
+
See the License for the specific language governing permissions and
|
| 93 |
+
limitations under the License.
|
| 94 |
+
```
|