Update README.md

README.md

@@ -1,36 +1,35 @@
 ---
 license: mit
 base_model:
-- inclusionAI/Ling-flash-base-2.0
+  - inclusionAI/Ling-flash-base-2.0
 pipeline_tag: text-generation
 library_name: transformers
 ---
 
-
-
 <p align="center">
     <img src="https://mdn.alipayobjects.com/huamei_qa8qxu/afts/img/A*4QxcQrBlTiAAAAAAQXAAAAgAemJ7AQ/original" width="100"/>
 <p>
 
-<p align="center">🤗 <a href="https://huggingface.co/inclusionAI">Hugging Face</a>&nbsp&nbsp | &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/inclusionAI">ModelScope</a></p>
-
+<p align="center">🤗 <a href="https://huggingface.co/inclusionAI">Hugging Face</a>&nbsp&nbsp | &nbsp&nbsp🤖 <a href="https://modelscope.cn/organization/inclusionAI">ModelScope</a>&nbsp&nbsp | &nbsp&nbsp🐙 <a href="https://zenmux.ai/inclusionai/ling-flash-2.0">ChatNow</a></p>
 
 ## Introduction
 
-Today, __Ling-flash-2.0__ is officially open-sourced! 🚀
-Following the release of the __language model [Ling-mini-2.0](https://huggingface.co/inclusionAI/Ling-mini-2.0)__ and the __thinking model [Ring-mini-2.0](https://huggingface.co/inclusionAI/Ring-mini-2.0)__, we are now open-sourcing the third MoE LLM under the __Ling 2.0 architecture: Ling-flash-2.0__, a language model with __100B total parameters__ and __6.1B activated parameters (4.8B non-embedding)__.
-Trained on __20T+ tokens of high-quality data__, together with __supervised fine-tuning__ and __multi-stage reinforcement learning__, Ling-flash-2.0 achieves __SOTA performance among dense models under 40B parameters__, despite activating only ~6B parameters. Compared to MoE models with larger activation/total parameters, it also demonstrates strong competitiveness. Notably, it delivers outstanding performance in __complex reasoning, code generation, and frontend development__.
+Today, **Ling-flash-2.0** is officially open-sourced! 🚀
+Following the release of the **language model [Ling-mini-2.0](https://huggingface.co/inclusionAI/Ling-mini-2.0)** and the **thinking model [Ring-mini-2.0](https://huggingface.co/inclusionAI/Ring-mini-2.0)**, we are now open-sourcing the third MoE LLM under the **Ling 2.0 architecture: Ling-flash-2.0**, a language model with **100B total parameters** and **6.1B activated parameters (4.8B non-embedding)**.
+Trained on **20T+ tokens of high-quality data**, together with **supervised fine-tuning** and **multi-stage reinforcement learning**, Ling-flash-2.0 achieves **SOTA performance among dense models under 40B parameters**, despite activating only ~6B parameters. Compared to MoE models with larger activation/total parameters, it also demonstrates strong competitiveness. Notably, it delivers outstanding performance in **complex reasoning, code generation, and frontend development**.
 
 ### Powerful Complex Reasoning Abilities
 
 We conducted a comprehensive evaluation of Ling-flash-2.0’s reasoning capabilities, reporting strong results on representative benchmarks:
-* __Multi-disciplinary knowledge reasoning__: GPQA-Diamond, MMLU-Pro
-* __Advanced mathematical reasoning__: AIME 2025, Omni-MATH, OptMATH (advanced mathematical optimization tasks)
-* __Challenging code generation__: LiveCodeBench v6, CodeForces-Elo
-* __Logical reasoning__: KOR-Bench, ARC-Prize
-* __Key regulated industries (Finance, Healthcare)__: FinanceReasoning, HealthBench
 
-Compared with __dense models under 40B__ (e.g., Qwen3-32B-Non-Thinking, Seed-OSS-36B-Instruct (think budget=0)) and __larger-activation/total-parameter MoE models__ (e.g., Hunyuan-A13B-Instruct, GPT-OSS-120B/low), __Ling-flash-2.0__ demonstrates stronger complex reasoning power. Moreover, it shows high competitiveness on __creative tasks__ (Creative Writing v3).
+- **Multi-disciplinary knowledge reasoning**: GPQA-Diamond, MMLU-Pro
+- **Advanced mathematical reasoning**: AIME 2025, Omni-MATH, OptMATH (advanced mathematical optimization tasks)
+- **Challenging code generation**: LiveCodeBench v6, CodeForces-Elo
+- **Logical reasoning**: KOR-Bench, ARC-Prize
+- **Key regulated industries (Finance, Healthcare)**: FinanceReasoning, HealthBench
+
+Compared with **dense models under 40B** (e.g., Qwen3-32B-Non-Thinking, Seed-OSS-36B-Instruct (think budget=0)) and **larger-activation/total-parameter MoE models** (e.g., Hunyuan-A13B-Instruct, GPT-OSS-120B/low), **Ling-flash-2.0** demonstrates stronger complex reasoning power. Moreover, it shows high competitiveness on **creative tasks** (Creative Writing v3).
+
 <p align="center">
     <img src="https://mdn.alipayobjects.com/huamei_fi95qp/afts/img/zxAvQ7QtrAwAAAAAQqAAAAgADkZ7AQFr/fmt.webp"/>
 <p>
@@ -45,11 +44,11 @@ Compared with __dense models under 40B__ (e.g., Qwen3-32B-Non-Thinking, Seed-OSS
     <img src="https://mdn.alipayobjects.com/huamei_fi95qp/afts/img/fMdiQZqYKSAAAAAAVdAAAAgADkZ7AQFr/fmt.avif"/>
 <p>
 
-Guided by [Ling Scaling Laws](https://arxiv.org/abs/2507.17702), Ling 2.0 adopts a __1/32 activation-ratio MoE architecture__, optimized across multiple design choices: expert granularity, shared-expert ratio, attention balance, __aux-loss-free + sigmoid routing strategy__, MTP layers, QK-Norm, Partial-RoPE, and more. These refinements enable __small-activation MoE__ models to achieve __7× efficiency gains__ over equivalent dense architectures.
-In other words, with just __6.1B activated parameters (4.8B non-embedding)__, __Ling-flash-2.0__ can match the performance of ~40B dense models. Thanks to its small activation size, it also delivers major inference speed advantages:
-* On __H20 hardware__, Ling-flash-2.0 achieves __200+ tokens/s__, offering __3× speedups__ compared to 36B dense models in everyday use.
-* With __YaRN extrapolation__, it supports __128K context length__, and as output length grows, its relative speedup can reach __7× or more__.
+Guided by [Ling Scaling Laws](https://arxiv.org/abs/2507.17702), Ling 2.0 adopts a **1/32 activation-ratio MoE architecture**, optimized across multiple design choices: expert granularity, shared-expert ratio, attention balance, **aux-loss-free + sigmoid routing strategy**, MTP layers, QK-Norm, Partial-RoPE, and more. These refinements enable **small-activation MoE** models to achieve **7× efficiency gains** over equivalent dense architectures.
+In other words, with just **6.1B activated parameters (4.8B non-embedding)**, **Ling-flash-2.0** can match the performance of ~40B dense models. Thanks to its small activation size, it also delivers major inference speed advantages:
 
+- On **H20 hardware**, Ling-flash-2.0 achieves **200+ tokens/s**, offering **3× speedups** compared to 36B dense models in everyday use.
+- With **YaRN extrapolation**, it supports **128K context length**, and as output length grows, its relative speedup can reach **7× or more**.
 
 <p align="center">
     <img src="https://mdn.alipayobjects.com/huamei_fi95qp/afts/img/oR9UTY7S0QgAAAAAgKAAAAgADkZ7AQFr/original"/>
@@ -59,25 +58,57 @@ In other words, with just __6.1B activated parameters (4.8B non-embedding)__, __
     <img src="https://mdn.alipayobjects.com/huamei_fi95qp/afts/img/Hid1RrgsCUAAAAAAQYAAAAgADkZ7AQFr/fmt.webp"/>
 <p>
 
-
 ## Model Downloads
 
 You can download the following table to see the various stage of Ling-flash-2.0 models. If you are located in mainland China, we also provide the model on ModelScope.cn to speed up the download process.
 
 <center>
 
-|       **Model**        | **Context Length** |                                                                             **Download**                                                                             |
-|:----------------------:| :----------------: |:--------------------------------------------------------------------------------------------------------------------------------------------------------------------:|
-|   Ling-flash-base-2.0   |        32K -> 128K (YaRN)         |     [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ling-flash-base-2.0) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ling-flash-base-2.0)     |
-|     Ling-flash-2.0      |        32K -> 128K (YaRN)        |          [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ling-flash-2.0) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ling-flash-2.0)          |
+|      **Model**      | **Context Length** |                                                                          **Download**                                                                          |
+| :-----------------: | :----------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------: |
+| Ling-flash-base-2.0 | 32K -> 128K (YaRN) | [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ling-flash-base-2.0) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ling-flash-base-2.0) |
+|   Ling-flash-2.0    | 32K -> 128K (YaRN) |      [🤗 HuggingFace](https://huggingface.co/inclusionAI/Ling-flash-2.0) <br>[🤖 ModelScope](https://www.modelscope.cn/models/inclusionAI/Ling-flash-2.0)      |
 
 </center>
 
 Note: If you are interested in previous version, please visit the past model collections in [Huggingface](https://huggingface.co/inclusionAI) or [ModelScope](https://modelscope.cn/organization/inclusionAI).
 
-
 ## Quickstart
 
+### 🚀 Try Online
+
+You can experience Ling-flash-2.0 online at: [ZenMux](https://zenmux.ai/inclusionai/ling-flash-2.0)
+
+### 🔌 API Usage
+
+You can also use Ling-flash-2.0 through API calls:
+
+```python
+from openai import OpenAI
+
+# 1. Initialize the OpenAI client
+client = OpenAI(
+    # 2. Point the base URL to the ZenMux endpoint
+    base_url="https://zenmux.ai/api/v1",
+    # 3. Replace with the API Key from your ZenMux user console
+    api_key="<your ZENMUX_API_KEY>",
+)
+
+# 4. Make a request
+completion = client.chat.completions.create(
+    # 5. Specify the model to use in the format "provider/model-name"
+    model="inclusionai/ling-flash-2.0",
+    messages=[
+        {
+            "role": "user",
+            "content": "What is the meaning of life?"
+        }
+    ]
+)
+
+print(completion.choices[0].message.content)
+```
+
 ### 🤗 Hugging Face Transformers
 
 Here is a code snippet to show you how to use the chat model with `transformers`:
@@ -177,7 +208,9 @@ vllm serve inclusionAI/Ling-flash-2.0 \
 ```
 
 To handle long context in vLLM using YaRN, we need to follow these two steps:
+
 1. Add a `rope_scaling` field to the model's `config.json` file, for example:
+
 ```json
 {
   ...,
@@ -188,24 +221,29 @@ To handle long context in vLLM using YaRN, we need to follow these two steps:
   }
 }
 ```
+
 2. Use an additional parameter `--max-model-len` to specify the desired maximum context length when starting the vLLM service.
 
 For detailed guidance, please refer to the vLLM [`instructions`](https://docs.vllm.ai/en/latest/).
 
-
 ### SGLang
 
 #### Environment Preparation
 
 We will later submit our model to SGLang official release, now we can prepare the environment following steps:
+
 ```shell
 pip3 install sglang==0.5.2rc0 sgl-kernel==0.3.7.post1
 ```
+
 You can use docker image as well:
+
 ```shell
 docker pull lmsysorg/sglang:v0.5.2rc0-cu126
 ```
+
 Then you should apply patch to sglang installation:
+
 ```shell
 # patch command is needed, run `yum install -y patch` if needed
 patch -d `python -c 'import sglang;import os; print(os.path.dirname(sglang.__file__))'` -p3 < inference/sglang/bailing_moe_v2.patch
@@ -213,9 +251,10 @@ patch -d `python -c 'import sglang;import os; print(os.path.dirname(sglang.__fil
 
 #### Run Inference
 
-BF16 and FP8 models are supported by SGLang now, it depends on the dtype of the model in ${MODEL_PATH}. They both share the same command in the following:  
+BF16 and FP8 models are supported by SGLang now, it depends on the dtype of the model in ${MODEL_PATH}. They both share the same command in the following:
 
 - Start server:
+
 ```shell
 python -m sglang.launch_server \
     --model-path $MODLE_PATH \
@@ -223,6 +262,7 @@ python -m sglang.launch_server \
     --trust-remote-code \
     --attention-backend fa3
 ```
+
 MTP is supported for base model, and not yet for chat model. You can add parameter `--speculative-algorithm NEXTN`
 to start command.
 
@@ -236,8 +276,6 @@ curl -s http://localhost:${PORT}/v1/chat/completions \
 
 More usage can be found [here](https://docs.sglang.ai/basic_usage/send_request.html)
 
-
-
 ### Finetuning
 
 We recommend you to use [Llama-Factory](https://github.com/hiyouga/LLaMA-Factory) to [finetune Ling](https://github.com/inclusionAI/Ling-V2/blob/main/docs/llamafactory_finetuning.md).
@@ -245,5 +283,3 @@ We recommend you to use [Llama-Factory](https://github.com/hiyouga/LLaMA-Factory
 ## License
 
 This code repository is licensed under [the MIT License](https://github.com/inclusionAI/Ling-V2/blob/master/LICENCE).
-
-