| { | |
| "method": "HLWQ", | |
| "weight_bits": 5, | |
| "kv_cache_bits": 3, | |
| "block_size": 128, | |
| "head_dim": 128, | |
| "num_layers": 64, | |
| "quantized_layers": 497, | |
| "skipped_layers": 0, | |
| "int4_group_size": 128, | |
| "vram_gb": 19.1, | |
| "base_model": "huihui-ai/Huihui-Qwopus3.5-27B-v3-abliterated", | |
| "benchmark": { | |
| "FP16": { | |
| "tok_s": 21.804779849809538, | |
| "peak_gb": 70.363202048, | |
| "time": 4.586150407791138 | |
| }, | |
| "Q3": { | |
| "tok_s": 22.092763850247692, | |
| "peak_gb": 19.130801664, | |
| "time": 4.526368935902913 | |
| }, | |
| "Q2": { | |
| "tok_s": 22.038845106697618, | |
| "peak_gb": 19.130801664, | |
| "time": 4.537442843119304 | |
| } | |
| } | |
| } |