MCPcopy
hub / github.com/Andyyyy64/whichllm / GGUFVariant

Class GGUFVariant

src/whichllm/models/types.py:7–10  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

5
6@dataclass
7class GGUFVariant:
8 filename: str
9 quant_type: str # "Q4_K_M", "Q8_0" etc
10 file_size_bytes: int
11
12
13@dataclass

Callers 15

display_planFunction · 0.90
display_plan_jsonFunction · 0.90
_synthesize_variants_for_official_repoFunction · 0.90
_parse_modelFunction · 0.90
dicts_to_modelsFunction · 0.90
test_estimate_vram_gguf_variantFunction · 0.90
test_estimate_vram_increases_with_contextFunction · 0.90
test_estimate_vram_small_modelFunction · 0.90
_make_variantFunction · 0.90
test_gddr6_estimate_scales_with_bandwidth_and_matches_measuredFunction · 0.90
_ggufFunction · 0.90
test_qwen3_next_strix_halo_matches_reported_generation_speedMethod · 0.90

Calls

no outgoing calls

Tested by 15

test_estimate_vram_gguf_variantFunction · 0.72
test_estimate_vram_increases_with_contextFunction · 0.72
test_estimate_vram_small_modelFunction · 0.72
_make_variantFunction · 0.72
test_gddr6_estimate_scales_with_bandwidth_and_matches_measuredFunction · 0.72
_ggufFunction · 0.72
test_qwen3_next_strix_halo_matches_reported_generation_speedMethod · 0.72
test_unknown_ultra_sparse_moe_uses_active_params_on_strix_haloMethod · 0.72
test_qwen3_30b_a3b_strix_halo_no_longer_uses_legacy_floorMethod · 0.72
test_strix_halo_moe_speed_is_medium_confidence_with_rangeMethod · 0.72
test_pick_gguf_variant_by_preferenceFunction · 0.72
test_pick_gguf_variant_with_filterFunction · 0.72