SVDQuant: a new 4-bit post-training quantization paradigm for diffusion models
The rapid scaling of diffusion models has created challenges in memory usage and latency, making them difficult to implement, particularly ...
Tag: posttraining
Meet BiLLM: a novel post-training binary quantization method designed specifically to compress pre-trained LLMs
Pretrained large language models (LLMs) have remarkable language processing capabilities, but require substantial computational resources. Binarization, which reduces model weights ...
Recommended.
