Text Generation
Transformers
Safetensors
qwen3
llama-factory
full
Generated from Trainer
conversational
text-generation-inference
Instructions to use ayh015/myLightningOPD with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ayh015/myLightningOPD with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ayh015/myLightningOPD") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("ayh015/myLightningOPD") model = AutoModelForCausalLM.from_pretrained("ayh015/myLightningOPD", device_map="auto") messages = [ {"role": "user", "content": "Who are you?"}, ] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, tokenize=True, return_dict=True, return_tensors="pt", ).to(model.device) outputs = model.generate(**inputs, max_new_tokens=40) print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use ayh015/myLightningOPD with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ayh015/myLightningOPD" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ayh015/myLightningOPD", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/ayh015/myLightningOPD
- SGLang
How to use ayh015/myLightningOPD with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "ayh015/myLightningOPD" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ayh015/myLightningOPD", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "ayh015/myLightningOPD" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ayh015/myLightningOPD", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use ayh015/myLightningOPD with Docker Model Runner:
docker model run hf.co/ayh015/myLightningOPD
Rollout Buffer
概述
Rollout Buffer 是用于辅助纯异步 agent 训练的独立组件,其主要功能是使用 slime 训练启动的 LLM OpenAI Server 进行智能体轨迹的生成。
工作流程
slime Training Process ←─── HTTP API ───→ Rollout Buffer
↓ ↓
LLM Server ←─────── HTTP Requests ─────── Agent Framework
↓ ↓
Model Response ──────────────────────→ Trajectory Generation
对于每一个不同的 Agent 任务,都应该对应一个独立的 Generator 类,负责生成该类任务的轨迹。Rollout Buffer 会自动读取并加载不同类型的 Generator。
快速开始
基本使用流程
- 复制模板:将
base_generator.py作为模板进行复制 - 修改任务类型:将
TASK_TYPE修改为您的任务名称(不能与其他 Generator 重复) - 实现核心函数:实现
run_rollout()函数 - 可选定制:根据需要重写五个可选函数
Generator 文件必须以 _generator.py 结尾,并放置在 generator/ 目录下:
generator/
├── base_generator.py # Math 任务实现(默认模板)
└── your_task_generator.py # 您的自定义任务
每个 Generator 文件必须定义 TASK_TYPE 与 run_rollout()。
此外,Rollout Buffer 还提供了一些可自定义的函数来满足不同任务的特殊需求。如果不提供自定义实现,系统将使用默认实现(位于 slime_plugins/rollout_buffer/default_func.py)。
示例脚本
请仿照 示例:Qwen3-4B 模型 文档中配置好 slime 的运行环境,下载数据,并转换模型 ckpt。之后分别运行
cd slime_plugins/rollout_buffer
bash rollout_buffer_example.sh
# In a different terminal
python buffer.py