Tag

LightRAG

A graph-based RAG framework pairing a knowledge graph with vector retrieval and a dual-level (low/high) query mode. New documents merge into the graph via set operations instead of triggering a rebuild, cutting the cost of keeping the index current.

AI Toolkit

Trains and fine-tunes diffusion models on consumer GPUs: LoRA and LoKr for image families like FLUX.1/2, SDXL and Qwen-Image, plus video models such as Wan 2.x and LTX. Layer-specific targeting, configurable VRAM, and a browser dashboard for runs.

Weights & Biases

Tracks every ML run — hyperparameters, metrics, checkpoints, dataset versions — into one dashboard you share as a live report, with Sweeps for tuning and a model registry. Weave extends it to LLM apps: tracing, evals, and production monitoring.

OpenVINO

Converts, quantizes, and runs deep learning models from PyTorch, TensorFlow, ONNX, and PaddlePaddle across Intel CPUs, GPUs, and NPUs without the training framework. Adds a GenAI pipeline for LLMs plus Hugging Face, vLLM, and LangChain integrations.

Transformers

Provides unified model definitions and a single API for pretrained text, vision, audio, and multimodal models for both training and inference. Emphasizes cross-framework compatibility (PyTorch/TF/JAX), pipeline-based inference, and direct access to 1M+ Hub checkpoints.

Hugging Face — Transformers

Turns model definitions into a shared layer across training and inference stacks, covering text, vision, audio, video, and multimodal models. Pipelines, Trainer, and generation APIs make pretrained models usable without locking teams to one framework.

PyTorch Image Models (timm)

Bundles hundreds of pretrained image backbones — ResNet, EfficientNet, ViT, ConvNeXt, Swin and more — behind one consistent API for classification and feature extraction, with training and inference scripts that reproduce published ImageNet results.

CleanRL (Clean Implementation of RL Algorithms)

Implements deep RL algorithms (PPO, DQN, SAC, TD3, DDPG, C51, PPG) as standalone single-file scripts — the PPO Atari variant is ~340 readable lines. Built for research debugging and reproducibility, with W&B and TensorBoard tracking.

NVIDIA NeMo

Build, fine-tune, and deploy speech AI on NVIDIA GPUs: ASR, text-to-speech, and speech LLMs in one PyTorch stack. Ships pretrained Parakeet/Canary recognition and Magpie TTS checkpoints; broader LLM/multimodal training now lives in v2.7.0.

NVIDIA NeMo Speech

Provides a toolkit and codebase for building, training, and deploying speech and multimodal models — Automatic Speech Recognition, Text-to-Speech, and speech-aware LLMs — with modular neural components and pre-trained checkpoints for PyTorch. Supports streaming/low-latency inference, multi-language models, and optional compiled kernels for acceleration.

PaddleOCR

Turns PDFs and images into structured Markdown or JSON for LLM use, recognizing text, tables, formulas, seals, and handwriting. Models scale from a 1.5M-param edge variant to a 0.9B vision-language model spanning 100+ languages.

Language Model Evaluation Harness

Unified framework for few-shot evaluation of generative language models across 60+ academic benchmarks. Supports multiple model backends (Hugging Face, vLLM, APIs, local servers), configurable prompts/YAML configs, and reproducible exports for leaderboards and research comparisons.