Tutorials
Scenario-based end-to-end tutorials — copy-paste runnable, with time estimate and difficulty
Each tutorial is a self-contained walkthrough. Open one and follow it top-to-bottom — copy-paste runs.
Training & Fine-tuning
LoRA fine-tune with LLaMA Factory on CCI
Install LLaMA Factory in CCI and run a LoRA SFT pass from the WebUI (Llama-3-8B-Instruct as the example).
PyTorch binary classifier on CCI
Train a small classifier on CCI with Jupyter + PyTorch — from env setup to plotting decision boundaries.
Inference & Serving
OpenAI-compatible inference with vLLM on CCI
Install vLLM in CCI, serve an OpenAI-compatible inference endpoint and test a completion (Baichuan2-7B-Chat as the example).
Application integration
Dify + RAG knowledge-base bot
Wire enterprise business data into Dify and build a RAG-backed knowledge-base Q&A bot.
LLaMA Factory Fine-tuning Guide
Use Aladdin IDE plugin + LLaMA Factory for single-GPU, multi-GPU single-node, and multi-node model fine-tuning.
Tooling & Environment
Linux cheat sheet for GPU pods
The ~40 commands you actually use after SSH-ing into a pod — files / permissions / SCP / nvidia-smi GPU triage / killing squatter processes.
Hugging Face mirror
Set HF_ENDPOINT to the regional mirror — three download methods (snapshot_download, huggingface-cli, from_pretrained).
查看其余 5 个 →Show 5 more →
Remote Development
Last updated on