twinkle
0.4.0.dev0
Usage Guide
Quick Start
Twinkle Installation
Server and Client
NPU (Ascend) Quick Start Guide
Twinkle Training Service on ModelScope
Qwen3.5 Training Best Practices
Components
Dataset
Data Format
Template
Preprocessor and Filter
Data Loading
Task Processor
Model
Sampler
Reward
Advantage
Advantage
GRPOAdvantage
RLOOAdvantage
Gym
Hub
Checkpoint Engine
Metrics
Loss
Loss Scale
LRScheduler
Patch
Plugin
Kernel
Training Middleware
twinkle
Advantage
View page source
Advantage
Advantage
GRPOAdvantage
RLOOAdvantage