huggingface/picotron
minimal 4D-parallelism framework for distributed LLM training and experimentation
View on index · View in 3D Map
Related repos
- Platform for automatic algorithm design using LLMs
- Orchestrates parallel function calling for LLMs
- Programming language for collaborating with AI language models
- Guide to fine-tuning LLMs with PyTorch and Hugging Face
- framework for fast, resource-efficient LLM inference on local hardware
- LLM training in C and CUDA, minimal dependencies
- Nanotron: minimalistic library for pretraining transformer models with simple API and scalable parallelism
- massively parallel programming language for GPUs and CPUs
- Toolkit for distributed training and inference of large AI models using parallelism and memory management techniques
- OpenAI-compatible proxy that boosts LLM reasoning accuracy with 20+ optimization techniques, no retraining needed
