fix docs

Author: tastelikefeet

README.md

@@ -81,7 +81,7 @@ pip install -e .
 
 We are rolling out training service built atop Twinkle✨ on ModelScope. It is currently in _Beta_. You may
 sign up for free access by joining the [Twinkle-Explorers](https://modelscope.cn/organization/twinkle-explorers) organization, and
-train via API endpoint  `base_url=https://www.modelscope.cn/twinkle`. For more details, please refer to 
+train via API endpoint  `base_url=https://www.modelscope.cn/twinkle`. For more details, please refer to
 our [documentation](docs/source_en/Usage%20Guide/ModelScope-Official-Resources.md).
 
 ## Supported Hardware
@@ -96,9 +96,9 @@ our [documentation](docs/source_en/Usage%20Guide/ModelScope-Official-Resources.m
 ## Supported Models
 
 We will be adding support for more models as new models are released. The following table lists current models
-supported on Twinkle✨ framework. 
+supported on Twinkle✨ framework.
 
->[!Note] 
+>[!Note]
 > For serverless training service accessed via `base_url=https://www.modelscope.cn/twinkle`, it currently supports
 > one training base at a time, and currently it is [Qwen3-30B-A3B-Instruct-2507](https://modelscope.cn/models/Qwen/Qwen3-30B-A3B-Instruct-2507).
 
@@ -203,7 +203,7 @@ from twinkle.dataset import Dataset, DatasetMeta
 from twinkle.preprocessor import SelfCognitionProcessor
 from twinkle.server.tinker.common import input_feature_to_datum
 
-base_model = 'ms://Qwen/Qwen2.5-0.5B-Instruct'
+base_model = 'ms://Qwen/Qwen3-30B-A3B-Instruct-2507'
 base_url='http://www.modelscope.cn/twinkle'
 api_key=os.environ.get('MODELSCOPE_TOKEN')
 

README_ZH.md

@@ -1,17 +1,11 @@
-<div align="center">
-
 ## ✨ What is Twinkle?
 
 A component library for large model training. Based on PyTorch, simpler, more flexible, production-ready.
 
-<p align="center">
 🧩 <b>Loosely Coupled Architecture</b> · Standardized Interfaces<br>
 🚀 <b>Multiple Runtime Modes</b> · torchrun / Ray / HTTP<br>
 🔌 <b>Multi-Framework Compatible</b> · Transformers / Megatron<br>
 👥 <b>Multi-Tenant Support</b> · Single Base Model Deployment
-</p>
-
-</div>
 
 ## Twinkle Compatibility
 
@@ -23,7 +17,6 @@ Twinkle and [ms-swift](https://github.com/modelscope/ms-swift) are both model tr
 - If you are a large model researcher who wants to customize models or training methods
 - If you are good at writing training loops and want to customize the training process
 - If you want to provide enterprise-level or commercial training platforms
-- If you lack training hardware and want to use community resources
 
 ### When to Choose ms-swift
 

docs/source_en/Usage Guide/ModelScope-Official-Resources.md

@@ -0,0 +1,34 @@
+# Twinkle Training Service on ModelScope
+
+Alongside the open-source release of the Twinkle framework, we also provide a hosted model training service (Training as a Service) powered by ModelScope's backend infrastructure. Developers can use this service to experience Twinkle's training API for free.
+
+The model currently running on the cluster is [Qwen/Qwen3-30B-A3B-Instruct-2507](https://www.modelscope.cn/models/Qwen/Qwen3-30B-A3B-Instruct-2507). Below are the detailed usage instructions:
+
+## Step 1. Register a ModelScope Account and Apply to Join the twinkle-explorers Organization
+
+Developers first need to register as a ModelScope user and apply to join the [Twinkle-Explorers](https://modelscope.cn/organization/twinkle-explorers) organization to obtain access permissions. The current free Serverless training experience is still in beta testing and is only available to users within the organization. You can also use Twinkle✨ by deploying the service locally.
+
+Registration link: https://www.modelscope.cn/
+
+After registering and being approved to join the [Twinkle-Explorers](https://modelscope.cn/organization/twinkle-explorers) organization, obtain your API-Key (i.e., the ModelScope platform access token) from this page: https://www.modelscope.cn/my/access/token.
+
+API endpoint: `base_url="https://www.modelscope.cn/twinkle"`
+
+## Step 2. Review the Cookbook and Customize Development
+
+We strongly recommend that developers review our [cookbook](https://github.com/modelscope/twinkle/tree/main/cookbook/client/) and build upon the training code provided there.
+
+Developers can customize datasets, advantage functions, rewards, templates, and more. However, the Loss component is not currently customizable since it needs to be executed on the server side (for security reasons). If you need support for additional Loss functions, you can upload your Loss implementation to ModelHub and contact us via the Q&A group or through an issue to have the corresponding component added to the whitelist.
+
+## Appendix: Supported Training Methods
+
+This model is a text-only model, so multimodal tasks are not currently supported. For text-only tasks, you can train using:
+
+1. Standard PT/SFT training methods, including Agentic training
+2. Self-sampling RL algorithms such as GRPO/RLOO
+3. Distillation methods like GKD/On-policy. Since the official ModelScope endpoint only supports a single model, the other Teacher/Student model must be prepared by the developer
+
+The current official environment only supports LoRA training, with the following requirements:
+
+1. Maximum rank = 32
+2. modules_to_save is not supported

docs/source_en/Usage Guide/Quick-Start.md

@@ -13,6 +13,7 @@ Twinkle DOCUMENTATION
    Usage Guide/Installation.md
    Usage Guide/Server and Client/index.rst
    Usage Guide/NPU-Support.md
+   Usage Guide/Train-as-a-Service.md
 
 .. toctree::
    :maxdepth: 2

docs/source_en/Usage Guide/Train-as-a-Service.md

@@ -13,6 +13,7 @@ Twinkle DOCUMENTATION
    使用指引/安装.md
    使用指引/服务端和客户端/index.rst
    使用指引/NPU的支持.md
+   使用指引/训练服务.md
 
 .. toctree::
    :maxdepth: 2