Update README.md
Browse files
README.md
CHANGED
|
@@ -33,12 +33,10 @@ pipeline_tag: text-generation
|
|
| 33 |
The Qwen3-Coder-30B model shows specialized strength in mathematical reasoning tasks (GSM-8K) where it significantly outperforms the larger models. However, it lags behind the 235B models in general knowledge (MMLU), commonsense reasoning (Hellaswag), and pronoun disambiguation (Winogrande). This suggests the 30B model may have been optimized specifically for coding and mathematical tasks at the expense of some general capabilities.
|
| 34 |
|
| 35 |
|
|
|
|
| 36 |
|
| 37 |
|
| 38 |
# Qwen3-Coder-30B-A3B-Instruct
|
| 39 |
-
<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
|
| 40 |
-
<img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
|
| 41 |
-
</a>
|
| 42 |
|
| 43 |
## Highlights
|
| 44 |
|
|
|
|
| 33 |
The Qwen3-Coder-30B model shows specialized strength in mathematical reasoning tasks (GSM-8K) where it significantly outperforms the larger models. However, it lags behind the 235B models in general knowledge (MMLU), commonsense reasoning (Hellaswag), and pronoun disambiguation (Winogrande). This suggests the 30B model may have been optimized specifically for coding and mathematical tasks at the expense of some general capabilities.
|
| 34 |
|
| 35 |
|
| 36 |
+
-----
|
| 37 |
|
| 38 |
|
| 39 |
# Qwen3-Coder-30B-A3B-Instruct
|
|
|
|
|
|
|
|
|
|
| 40 |
|
| 41 |
## Highlights
|
| 42 |
|