All authors
Claude Skills by nota-america
github.com/nota-america1,679 skills4 installs2,242 views
- LitgptImplements and trains LLMs using Lightning AI's LitGPT with 20+Votes: 0GitHub stars: 81
- MambaState-space model with O(n) complexity vs Transformers' O(n²). 5×Votes: 0GitHub stars: 81
- NanogptEducational GPT implementation in ~300 lines. Reproduces GPT-2Votes: 0GitHub stars: 81
- RwkvRNN+Transformer hybrid with O(n) inference. Linear time, infiniteVotes: 0GitHub stars: 81
- TorchtitanProvides PyTorch-native distributed LLM pretraining usingVotes: 0GitHub stars: 81
- Huggingface TokenizersFast tokenizers optimized for research and production. Rust-basedVotes: 0GitHub stars: 81
- SentencepieceLanguage-independent tokenizer treating text as raw Unicode.Votes: 0GitHub stars: 81
- AxolotlExpert guidance for fine-tuning LLMs with Axolotl - YAML configs,Votes: 0GitHub stars: 81
- Llama FactoryExpert guidance for fine-tuning LLMs with LLaMA-Factory - WebUIVotes: 0GitHub stars: 81
- PeftParameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+Votes: 0GitHub stars: 81
- UnslothExpert guidance for fast fine-tuning with Unsloth - 2-5x fasterVotes: 0GitHub stars: 81
- NnsightProvides guidance for interpreting and manipulating neural networkVotes: 0GitHub stars: 81
- PyveneProvides guidance for performing causal interventions on PyTorchVotes: 0GitHub stars: 81
- SaelensProvides guidance for training and analyzing Sparse AutoencodersVotes: 0GitHub stars: 81
- Transformer LensProvides guidance for mechanistic interpretability research usingVotes: 0GitHub stars: 81
- Nemo CuratorGPU-accelerated data curation for LLM training. SupportsVotes: 0GitHub stars: 81
- Ray DataScalable data processing for ML workloads. Streaming executionVotes: 0GitHub stars: 81
- Grpo Rl TrainingExpert guidance for GRPO/RL fine-tuning with TRL for reasoning andVotes: 0GitHub stars: 81
- MilesProvides guidance for enterprise-grade RL training using miles, aVotes: 0GitHub stars: 81
- OpenrlhfHigh-performance RLHF framework with Ray+vLLM acceleration. Use forVotes: 0GitHub stars: 81
- SimpoSimple Preference Optimization for LLM alignment. Reference-freeVotes: 0GitHub stars: 81
- SlimeProvides guidance for LLM post-training with RL using slime, aVotes: 0GitHub stars: 81
- TorchforgeProvides guidance for PyTorch-native agentic RL using torchforge,Votes: 0GitHub stars: 81
- Trl Fine TuningFine-tune LLMs using reinforcement learning with TRL - SFT forVotes: 0GitHub stars: 81
- VerlProvides guidance for training LLMs with reinforcement learningVotes: 0GitHub stars: 81
- Constitutional AiAnthropic's method for training harmless AI throughVotes: 0GitHub stars: 81
- LlamaguardMeta's 7-8B specialized moderation model for LLM input/outputVotes: 0GitHub stars: 81
- Nemo GuardrailsNVIDIA's runtime safety framework for LLM applications. FeaturesVotes: 0GitHub stars: 81
- Prompt GuardMeta's 86M prompt injection and jailbreak detector. FiltersVotes: 0GitHub stars: 81
- AccelerateSimplest distributed training API. 4 lines to add distributedVotes: 0GitHub stars: 81
- DeepspeedExpert guidance for distributed training with DeepSpeed - ZeROVotes: 0GitHub stars: 81
- Megatron CoreTrains large language models (2B-462B parameters) using NVIDIAVotes: 0GitHub stars: 81
- Pytorch Fsdp2Adds PyTorch FSDP2 (fully_shard) to training scripts with correctVotes: 0GitHub stars: 81
- Pytorch LightningHigh-level PyTorch framework with Trainer class, automaticVotes: 0GitHub stars: 81
- Ray TrainDistributed training orchestration across clusters. ScalesVotes: 0GitHub stars: 81
- Lambda LabsReserved and on-demand GPU cloud instances for ML training andVotes: 0GitHub stars: 81
- ModalServerless GPU cloud platform for running ML workloads. Use whenVotes: 0GitHub stars: 81
- SkypilotMulti-cloud orchestration for ML workloads with automatic costVotes: 0GitHub stars: 81
- AwqActivation-aware weight quantization for 4-bit LLM compression withVotes: 0GitHub stars: 81
- BitsandbytesQuantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction withVotes: 0GitHub stars: 81
- Flash AttentionOptimizes transformer attention with Flash Attention for 2-4xVotes: 0GitHub stars: 81
- GgufGGUF format and llama.cpp quantization for efficient CPU/GPUVotes: 0GitHub stars: 81
- GptqPost-training 4-bit quantization for LLMs with minimal accuracyVotes: 0GitHub stars: 81
- HqqHalf-Quadratic Quantization for LLMs without calibration data. UseVotes: 0GitHub stars: 81
- Ml Training RecipesBattle-tested PyTorch training recipes for all domains — LLMs,Votes: 0GitHub stars: 81
- Bigcode Evaluation HarnessEvaluates code generation models across HumanEval, MBPP, MultiPL-E,Votes: 0GitHub stars: 81
- Lm Evaluation HarnessEvaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval,Votes: 0GitHub stars: 81
- Nemo EvaluatorEvaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU,Votes: 0GitHub stars: 81
- Llama CppRuns LLM inference on CPU, Apple Silicon, and consumer GPUs withoutVotes: 0GitHub stars: 81
- SglangFast structured generation and serving for LLMs with RadixAttentionVotes: 0GitHub stars: 81