{ "format": { "8-bit layers": "uint8 blocks of 32 with fp32 scale + uint8 zero point, dequantized in-graph (Cast/Sub/Mul, folded at load)", "4-bit layers": "uint4 blocks of 32, MatMulNBits" }, "group_size": 32, "calib_pages": 256, "eval_pages": 400, "fp32_layers": [ "cnn.proj", "page_in.net.0", "page_in.net.3", "blocks.0.rel", "blocks.1.rel", "blocks.2.rel", "blocks.3.rel", "blocks.4.rel", "blocks.5.rel", "blocks.6.rel", "blocks.7.rel", "ctx_head" ], "int4_8bit_edge_layers": [ "tok_in.net.0", "tok_in.net.3", "img_in.net.0", "img_in.net.3", "color_head.net.0", "color_head.net.3" ], "variants": { "fp32": { "haiku_dE": 0.03649106619724383, "native_dE": 0.07405353200621903, "size_mb": 63.45, "cpu_ms_median": 9.31 }, "int8": { "dE_vs_fp32": 0.0001628809163776168, "dE_max": 0.02378411777317524, "ctx_max_abs": 0.013257980346679688, "haiku_dE": 0.036490150559693575, "native_dE": 0.074042251650244, "elem_decisions_same_as_fp32": 1.0, "pack_parity_max_abs": 1.7881393432617188e-06, "size_mb": 20.59, "cpu_ms_median": 9.77 }, "int4": { "dE_vs_fp32": 0.00039874050049547804, "dE_max": 0.10922706872224808, "ctx_max_abs": 0.07820296287536621, "haiku_dE": 0.03647379979180793, "native_dE": 0.07408187076449395, "elem_decisions_same_as_fp32": 0.9995768986672308, "pack_parity_max_abs": 2.9206275939941406e-06, "size_mb": 13.27, "cpu_ms_median": 8.29 } } }