Runs on a 16 GB GPU Collection GGUFs measured in Abyss with the GPU capped at 16 GB. Most need 64 GB RAM; see each card. • 7 items • Updated 1 day ago
Runs on a 12 GB GPU Collection GGUFs measured in Abyss with the GPU capped at 12 GB or less. Each card lists the RAM it needs and the speed. • 3 items • Updated about 22 hours ago
QSA long-context GGUF Collection Qwen-family GGUFs with a QSA sparse-attention indexer: long contexts decode faster. Abyss only. Dense with STRATA_Q35_QSA=0. • 2 items • Updated 1 day ago
Ornith-1.5 GGUF Collection Ornith-1.5 GGUFs (BETA). 16G64G = 16 GB GPU + 64 GB RAM. 122B: 16G64G, 24G64G. 397B: 32G128G, 16G64G; DGX Spark planned. Runs on Abyss. • 5 items • Updated 1 day ago
Runs on a 32 GB GPU Collection GGUFs measured in Abyss on an RTX 5090 (32 GB) with 64-128 GB RAM: GLM-5.3-Flash, DeepSeek V4 / V4.1, Ornith-1.5-397B. • 7 items • Updated 1 day ago
Runs on a 24 GB GPU Collection GGUFs measured in Abyss with the GPU capped at 24 GB, with 64 GB RAM. See each card for the speed. • 2 items • Updated 1 day ago
Runs on a 16 GB GPU Collection GGUFs measured in Abyss with the GPU capped at 16 GB. Most need 64 GB RAM; see each card. • 7 items • Updated 1 day ago