https://pypi.org/project/petallm/ Recent updates to the Python Package Index for petallm en https://pypi.org/project/petallm/2.11.0/ PetaLLM allows single 4GB GPU card to run 70B large language models without quantization, distillation or pruning. 8GB vmem to run 405B Llama3.1. gavinli@animaai.cloud Sun, 12 Apr 2026 03:16:52 GMT