Update README.md
#6
by
m1ngcheng
- opened
README.md
CHANGED
|
@@ -1,3 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
pipeline_tag: text-generation
|
|
@@ -43,14 +46,14 @@ In the **AIME 25** benchmark, Ling-1T extends the **Pareto frontier** of reasoni
|
|
| 43 |
|
| 44 |
Ling-1T excels in visual reasoning and front-end code generation tasks, combining deep semantic understanding with precise code synthesis.
|
| 45 |
We introduce a hybrid *Syntax–Function–Aesthetics* reward mechanism, enabling the model to not only generate correct and functional code but also demonstrate a refined sense of **visual aesthetics**.
|
| 46 |
-
On **ArtifactsBench**, Ling-1T ranks **first among open-source models**, and the benchmark visualizations in this card were, in fact, *generated by Ling-1T itself*.
|
| 47 |
|
| 48 |
|
| 49 |
### Emergent Intelligence at Trillion-Scale
|
| 50 |
|
| 51 |
Scaling to the trillion-parameter level has revealed strong **emergent reasoning and transfer capabilities**.
|
| 52 |
For example, in the **BFCL V3** tool-use benchmark, Ling-1T achieves **≈ 70% tool-call accuracy** with only light instruction tuning—despite having seen no large-scale trajectory data during training.
|
| 53 |
-
Ling-1T can:
|
| 54 |
|
| 55 |
* Interpret complex natural-language instructions
|
| 56 |
* Transform abstract logic into functional visual components
|
|
@@ -327,7 +330,7 @@ More usage can be found [here](https://docs.sglang.ai/basic_usage/send_request.h
|
|
| 327 |
|
| 328 |
## Limitations & Future Plans
|
| 329 |
|
| 330 |
-
While **Ling-1T** has made strong progress in efficient reasoning, cross-domain generalization, and training efficiency, several limitations remain:
|
| 331 |
|
| 332 |
* **GQA-based attention**: stable for long-context reasoning but relatively costly. Future versions will adopt **hybrid attention** to improve efficiency.
|
| 333 |
* **Limited agentic ability**: current model has room to grow in multi-turn interaction, long-term memory, and tool use.
|
|
|
|
| 1 |
+
|
| 2 |
+
|
| 3 |
+
|
| 4 |
---
|
| 5 |
license: mit
|
| 6 |
pipeline_tag: text-generation
|
|
|
|
| 46 |
|
| 47 |
Ling-1T excels in visual reasoning and front-end code generation tasks, combining deep semantic understanding with precise code synthesis.
|
| 48 |
We introduce a hybrid *Syntax–Function–Aesthetics* reward mechanism, enabling the model to not only generate correct and functional code but also demonstrate a refined sense of **visual aesthetics**.
|
| 49 |
+
On **ArtifactsBench**, [Ling-1T](https://zenmux.ai/inclusionai/ling-1t?utm_source=hf_inclusionAI) ranks **first among open-source models**, and the benchmark visualizations in this card were, in fact, *generated by Ling-1T itself*.
|
| 50 |
|
| 51 |
|
| 52 |
### Emergent Intelligence at Trillion-Scale
|
| 53 |
|
| 54 |
Scaling to the trillion-parameter level has revealed strong **emergent reasoning and transfer capabilities**.
|
| 55 |
For example, in the **BFCL V3** tool-use benchmark, Ling-1T achieves **≈ 70% tool-call accuracy** with only light instruction tuning—despite having seen no large-scale trajectory data during training.
|
| 56 |
+
[Ling-1T](https://zenmux.ai/inclusionai/ling-1t?utm_source=hf_inclusionAI) can:
|
| 57 |
|
| 58 |
* Interpret complex natural-language instructions
|
| 59 |
* Transform abstract logic into functional visual components
|
|
|
|
| 330 |
|
| 331 |
## Limitations & Future Plans
|
| 332 |
|
| 333 |
+
While **[Ling-1T](https://zenmux.ai/inclusionai/ling-1t?utm_source=hf_inclusionAI)** has made strong progress in efficient reasoning, cross-domain generalization, and training efficiency, several limitations remain:
|
| 334 |
|
| 335 |
* **GQA-based attention**: stable for long-context reasoning but relatively costly. Future versions will adopt **hybrid attention** to improve efficiency.
|
| 336 |
* **Limited agentic ability**: current model has room to grow in multi-turn interaction, long-term memory, and tool use.
|