MCPcopy Create free account
hub / github.com/Icecubesaad/OpenLLM-Studio / recommend_quantization

Method recommend_quantization

src-tauri/src/optimization.rs:35–41  ·  view source on GitHub ↗

Recommend quantization level based on VRAM

(vram_gb: f64)

Source from the content-addressed store, hash-verified

33
34 /// Recommend quantization level based on VRAM
35 fn recommend_quantization(vram_gb: f64) -> String {
36 match vram_gb {
37 v if v >= 8.0 => "Q5_K_M or Q8_0".to_string(),
38 v if v >= 4.0 => "Q4_K_M".to_string(),
39 _ => "Q2_K or Q3_K_S".to_string(),
40 }
41 }
42
43 /// Calculate batch size based on GPU backend
44 fn calculate_batch_size(gpu_backend: &GpuBackend) -> i32 {

Callers

nothing calls this directly

Calls 1

to_stringMethod · 0.80

Tested by

no test coverage detected