Agent Skills by Peijia Qin
Showing 112 vetted skills indexed across 1 GitHub repositories.
t2ancegguf-quantization
Quantize models to GGUF format for CPU inference with llama.cpp.
t2anceawq-quantization
Automate 4-bit AWQ quantization of large language models for GPU deployment.
t2ancehqq-quantization
Quantize large language model weights without calibration data.
t2ancegptq
Quantize large language models to 4-bit precision for memory reduction and faster inference.
t2anceoptimizing-attention-flash
Apply Flash Attention to transformer models for reduced memory usage and increased speed.
t2ancequantizing-models-bitsandbytes
Quantize LLMs to 8-bit or 4-bit with bitsandbytes for memory reduction.
t2ancemlflow
Track ML experiments and manage model lifecycles across training, registry, and deployment.
t2ancetensorboard
Visualize training metrics, graphs, and embeddings to analyze ML experiments.
t2anceweights-and-biases
Automate ML experiment tracking, visualization, and artifact management with Weights & Biases.
t2ancenemo-curator
Clean and curate multi-modal LLM training data with GPU-accelerated filtering and deduplication.
t2anceray-data
Distribute data transformations across CPU/GPU clusters with Ray Data.
t2anceserving-llms-vllm
Deploy and manage high-throughput LLM serving with vLLM-powered OpenAI-compatible endpoints.
t2ancetensorrt-llm
Deploy high-throughput LLM inference with TensorRT-LLM on NVIDIA GPUs.
t2ancesglang
Automate prompt prefix caching for LLM serving with SGLang.
t2ancellama-cpp
Run LLM inference on CPU and non-NVIDIA hardware with GGUF models.
t2ancehuggingface-tokenizers
Train and load HuggingFace tokenizers with Rust-based Python bindings.
t2ancesentencepiece
Creates SentencePiece subword tokenizers from raw corpora for multilingual NLP preprocessing.
t2ancellama-factory
Guides users to fine-tune LLMs via LLaMA-Factory WebUI with no-code configuration.
t2ancepeft-fine-tuning
Fine-tune 7B–70B language models using LoRA and QLoRA adapters.
t2anceaxolotl
Design and execute Axolotl-based fine-tuning workflows with YAML configurations.
t2anceunsloth
Guide memory-efficient LLM fine-tuning with Unsloth using LoRA and QLoRA workflows.
t2anceopenrlhf-training
Orchestrate distributed RLHF training for large language models with Ray and vLLM.
t2ancesimpo-training
Optimize LLM alignment with reference-free SimPO preference optimization.
t2anceslime-rl-training
Guide RL post-training of GLM-scale LLMs with Megatron-LM and SGLang workflows.