Hugging Face Jobs gets Async GRPO with LoRA

hugging face has announced the integration of async grpo with lora across its hf jobs platform. this update enables more efficient model training and tuning, particularly for large language models. the deepseek-ai/deepseek-r1-distill-qwen-1.5b model is one of the first to utilize this new capability. researchers and engineers working with similar models will benefit from this update. the integration streamlines the training process, reducing the need for nccl.