ar-llm-browser/ALLaM-7B-SparseGPT-50
7B • Updated • 6
ALLaM (7B) Compressed
Collection of compressed ALLaM (by humain-ai) models. Methods include multiple quantization formats and pruning with SparseGPT and Wanda.
-
-
ar-llm-browser/ALLaM-7B-SparseGPT-50-2of4
5B • Updated • 4 -
ar-llm-browser/ALLaM-7B-Wanda-50
7B • Updated • 5
ar-llm-browser/ALLaM-7B-int4
7B • Updated • 3Note QLoRA quantization (NormalFloat4)
ar-llm-browser/ALLaM-7B-int8
7B • Updated • 3Note LLM.int8() quantization (int8)