DoanNgocCuong
Community@doanngoccuong · Hanoi
NHẤT HƯỚNG BÁT CHÁNH ĐẠO
Agent Skills by DoanNgocCuong
Showing 162 vetted skills indexed across 2 GitHub repositories.
DoanNgocCuongdspy
Develop AI systems with declarative programming and automatic prompt optimization in DSPy.
DoanNgocCuongopenrlhf-training
Train large language models with PPO, GRPO, RLOO, and DPO algorithms using Ray and vLLM.
DoanNgocCuongnemo-guardrails
Implement runtime safety guardrails for LLM applications with NeMo Guardrails.
DoanNgocCuongllamaindex
Ingest documents from over 300 sources and query them with RAG.
DoanNgocCuongdeepspeed
Optimize large-scale deep learning training with DeepSpeed distributed strategies.
DoanNgocCuongcrewai-multi-agent
Orchestrate role-based AI agent teams for sequential or hierarchical task execution.
DoanNgocCuongmoe-training
Train Mixture of Experts models using DeepSpeed and HuggingFace Transformers.
DoanNgocCuongnnsight-remote-interpretability
Inspect and manipulate PyTorch neural network internals via nnsight and NDIF.
DoanNgocCuonglangsmith-observability
Trace, evaluate, and monitor LLM applications with the LangSmith SDK.
DoanNgocCuongaudiocraft-audio-generation
Generate music and sound effects from text using AudioCraft.
DoanNgocCuongnemo-curator
Accelerates LLM dataset curation with GPU-based deduplication, filtering, PII redaction, and NSFW detection.
DoanNgocCuonghuggingface-accelerate
Simplify distributed PyTorch training across GPUs, nodes, and TPUs.
DoanNgocCuongml-paper-writing
Draft academic papers from research repositories and verify citations via APIs.
DoanNgocCuongchroma
Manage vector embeddings and metadata for semantic search and RAG.
DoanNgocCuongprompt-guard
Detect prompt injections and jailbreak attempts in LLM applications.
DoanNgocCuongevaluating-llms-harness
Evaluate LLMs against over 60 academic benchmarks like MMLU and GSM8K.
DoanNgocCuongllama-factory
Provides guidance for fine-tuning LLMs with LLaMA-Factory via CLI or WebUI.
DoanNgocCuongknowledge-distillation
Compress large language models into smaller student models using knowledge distillation techniques.
DoanNgocCuongsimpo-training
Fine-tune LLMs with reference-free preference optimization using the SimPO algorithm.
DoanNgocCuongslime-rl-training
Run GRPO training jobs for Qwen3-4B with Megatron-LM and SGLang.
DoanNgocCuongblip-2-vision-language
Generate image captions and answer visual questions with BLIP-2.
DoanNgocCuongpytorch-fsdp2
Integrate PyTorch FSDP2 fully_shard into training scripts with DTensor sharding and DCP checkpointing.
DoanNgocCuongimplementing-llms-litgpt
Trains and fine-tunes LLMs with LitGPT using LoRA/QLoRA.
DoanNgocCuongsegment-anything-model
Segment objects in images via points, boxes, or automatic mask generation.