File size: 708 Bytes
163ace0
dc08a9b
 
 
db8d27a
dc08a9b
 
 
 
163ace0
 
0130e06
163ace0
dc08a9b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
---
base_model: Qwen/Qwen3-0.6B
datasets:
- LucidityAI/Astral-Post-Training-Dataset
tags:
- code
- chemistry
- finance
- biology
---

# Astral-0.6B-Flash

Astral 0.6B is the smallest sized model in the Astral family. It was fine-tuned from Qwen3 0.6b on LucidityAI/Astral-Post-Training-Dataset. 


As with usual Qwen3 models, reasoning can be toggled through the usage of ```/no_think``` or not.


### Example Prompt (ChatML Format (THINKING)):

```xml
<|im_start|>user
What is the capital of France?
<|im_end|>
<|im_start|>assistant
<think>
```

### Example Prompt (ChatML Format (NON-THINKING)):

```xml
<|im_start|>user
What is the capital of France? /no_think
<|im_end|>
<|im_start|>assistant
<think>
```