[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"summary-185c7e9786934be3-building-recurrent-depth-transformers-with-openmyt-summary":3,"summaries-facets-categories":125,"summary-related-185c7e9786934be3-building-recurrent-depth-transformers-with-openmyt-summary":4142},{"id":4,"title":5,"ai":6,"body":13,"categories":89,"created_at":91,"date_modified":91,"description":83,"extension":92,"faq":91,"featured":93,"kicker_label":91,"meta":94,"navigation":106,"path":107,"published_at":108,"question":91,"scraped_at":109,"seo":110,"sitemap":111,"source_id":112,"source_name":113,"source_type":114,"source_url":115,"stem":116,"tags":117,"thumbnail_url":91,"tldr":122,"tweet":91,"unknown_tags":123,"__hash__":124},"summaries\u002Fsummaries\u002F185c7e9786934be3-building-recurrent-depth-transformers-with-openmyt-summary.md","Building Recurrent-Depth Transformers with OpenMythos",{"provider":7,"model":8,"input_tokens":9,"output_tokens":10,"processing_time_ms":11,"cost_usd":12},"openrouter","google\u002Fgemini-3.1-flash-lite",10779,592,3177,0.00358275,{"type":14,"value":15,"toc":82},"minimark",[16,21,25,29,32,55,59,62],[17,18,20],"h2",{"id":19},"the-recurrent-depth-transformer-rdt-premise","The Recurrent-Depth Transformer (RDT) Premise",[22,23,24],"p",{},"OpenMythos introduces a framework for building recurrent-depth transformers that allow for dynamic compute scaling at inference time. Unlike standard feed-forward transformers where depth is fixed at training, RDTs enable the model to perform additional computation by looping through its layers. This allows a single, fixed-parameter model to trade inference latency for increased reasoning depth, effectively extending its problem-solving capacity on complex tasks without requiring additional training.",[17,26,28],{"id":27},"practical-implementation-and-architecture","Practical Implementation and Architecture",[22,30,31],{},"The OpenMythos library supports modern architectural components, including:",[33,34,35,43,49],"ul",{},[36,37,38,42],"li",{},[39,40,41],"strong",{},"Attention Variants",": Supports both Multi-Latent Attention (MLA), similar to DeepSeek-V2, for compressed KV caching, and Grouped-Query Attention (GQA).",[36,44,45,48],{},[39,46,47],{},"Sparse MoE",": Integrates Mixture-of-Experts components with shared experts to maintain parameter efficiency.",[36,50,51,54],{},[39,52,53],{},"Stability Monitoring",": The framework provides tools to calculate the spectral radius of the recurrent injection matrix. Maintaining a spectral radius (ρ) less than 1 is critical for ensuring the stability of the recurrent loops during training and inference.",[17,56,58],{"id":57},"evaluating-loop-scaled-reasoning","Evaluating Loop-Scaled Reasoning",[22,60,61],{},"The author demonstrates the RDT capability using a synthetic compositional reasoning task: predicting the sum of digit chains modulo 7.",[33,63,64,70,76],{},[36,65,66,69],{},[39,67,68],{},"Training Strategy",": The model is trained with a fixed number of recurrent loops (4) using AdamW and a cosine learning rate schedule. Loss is calculated specifically at the position following the 'EQ' token.",[36,71,72,75],{},[39,73,74],{},"Inference Scaling",": Once trained, the model's reasoning depth is tested by varying the number of loops (1, 2, 4, 6, 8) at inference time.",[36,77,78,81],{},[39,79,80],{},"Results",": The experiments show that increasing the loop count allows the model to maintain or improve accuracy on out-of-distribution (OOD) tasks involving longer digit chains. This confirms that the recurrent mechanism successfully enables the model to perform deeper reasoning by re-utilizing its existing weights, rather than relying on a static depth architecture.",{"title":83,"searchDepth":84,"depth":84,"links":85},"",2,[86,87,88],{"id":19,"depth":84,"text":20},{"id":27,"depth":84,"text":28},{"id":57,"depth":84,"text":58},[90],"AI & LLMs",null,"md",false,{"content_references":95,"triage":101},[96],{"type":97,"title":98,"url":99,"context":100},"tool","OpenMythos","https:\u002F\u002Fgithub.com\u002Fkyegomez\u002FOpenMythos","recommended",{"relevance":102,"novelty":103,"quality":102,"actionability":103,"composite":104,"reasoning":105},4,3,3.6,"Category: AI & LLMs. The article discusses a novel framework for building recurrent-depth transformers, which addresses a specific audience pain point regarding model efficiency and reasoning depth. It provides practical implementation details, but lacks a step-by-step guide for application.",true,"\u002Fsummaries\u002F185c7e9786934be3-building-recurrent-depth-transformers-with-openmyt-summary","2026-05-22 07:39:30","2026-05-22 11:00:25",{"title":5,"description":83},{"loc":107},"185c7e9786934be3","MarkTechPost","article","https:\u002F\u002Fwww.marktechpost.com\u002F2026\u002F05\u002F22\u002Fbuild-recurrent-depth-transformers-with-openmythos-for-mla-gqa-sparse-moe-and-loop-scaled-reasoning\u002F","summaries\u002F185c7e9786934be3-building-recurrent-depth-transformers-with-openmyt-summary",[118,119,120,121],"llm","python","ai-tools","transformers","OpenMythos enables recurrent-depth transformers that trade inference-time compute for deeper reasoning by reusing model parameters through recurrent loops.",[121],"62XmVzaYyL5oySv5gojtjSjiwB4kDYGOtTwdyU6StAM",[126,129,132,134,137,140,142,144,146,148,150,152,155,157,159,161,163,165,167,169,171,173,175,177,179,181,184,187,189,191,194,196,198,201,203,205,207,209,211,213,215,217,219,221,223,226,228,230,232,234,236,238,240,242,244,246,248,250,252,254,256,258,260,262,264,266,268,270,272,274,276,278,280,282,284,286,288,290,292,294,296,298,300,302,304,306,308,310,312,314,316,318,320,322,324,326,328,330,332,334,336,338,340,342,344,346,348,350,352,354,356,358,360,362,364,366,368,370,372,374,376,378,380,382,384,386,388,390,392,394,396,398,400,402,404,406,408,410,412,414,416,418,420,422,424,426,428,430,432,434,436,438,440,442,444,446,448,450,452,454,456,458,460,462,464,466,468,470,472,474,476,478,480,482,484,486,488,490,492,494,496,498,500,502,504,506,508,510,512,514,516,518,520,522,524,526,528,530,532,534,536,538,540,542,544,546,548,550,552,554,556,558,560,562,564,566,568,570,572,574,576,578,580,582,584,586,588,590,592,594,596,598,600,602,604,606,608,610,612,614,616,618,620,622,624,626,628,630,632,634,636,638,640,642,644,646,648,650,652,654,656,658,660,662,664,666,668,670,672,674,676,678,680,682,684,686,688,690,692,694,696,698,700,702,704,706,708,710,712,714,716,718,720,722,724,726,728,730,732,734,736,738,740,742,744,746,748,750,752,754,756,758,760,762,764,766,768,770,772,774,776,778,780,782,784,786,788,790,792,794,796,798,800,802,804,806,808,810,812,814,816,818,820,822,824,826,828,830,832,834,836,838,840,842,844,846,848,850,852,854,856,858,860,862,864,866,868,870,872,874,876,878,880,882,884,886,888,890,892,894,896,898,900,902,904,906,908,910,912,914,916,918,920,922,924,926,928,930,932,934,936,938,940,942,944,946,948,950,952,954,956,958,960,962,964,966,968,970,972,974,976,978,980,982,984,986,988,990,992,994,996,998,1000,1002,1004,1006,1008,1010,1012,1014,1016,1018,1020,1022,1024,1026,1028,1030,1032,1034,1036,1038,1040,1042,1044,1046,1048,1050,1052,1054,1056,1058,1060,1062,1064,1066,1068,1070,1072,1074,1076,1078,1080,1082,1084,1086,1088,1090,1092,1094,1096,1098,1100,1102,1104,1106,1108,1110,1112,1114,1116,1118,1120,1122,1124,1126,1128,1130,1132,1134,1136,1138,1140,1142,1144,1146,1148,1150,1152,1154,1156,1158,1160,1162,1164,1166,1168,1170,1172,1174,1176,1178,1180,1182,1184,1186,1188,1190,1192,1194,1196,1198,1200,1202,1204,1206,1208,1210,1212,1214,1216,1218,1220,1222,1224,1226,1228,1230,1232,1234,1236,1238,1240,1242,1244,1246,1248,1250,1252,1254,1256,1258,1260,1262,1264,1266,1268,1270,1272,1274,1276,1278,1280,1282,1284,1286,1288,1290,1292,1294,1296,1298,1300,1302,1304,1306,1308,1310,1312,1314,1316,1318,1320,1322,1324,1326,1328,1330,1332,1334,1336,1338,1340,1342,1344,1346,1348,1350,1352,1354,1356,1358,1360,1362,1364,1366,1368,1370,1372,1374,1376,1378,1380,1382,1384,1386,1388,1390,1392,1394,1396,1398,1400,1402,1404,1406,1408,1410,1412,1414,1416,1418,1420,1422,1424,1426,1428,1430,1432,1434,1436,1438,1440,1442,1444,1446,1448,1450,1452,1454,1456,1458,1460,1462,1464,1466,1468,1470,1472,1474,1476,1478,1480,1482,1484,1486,1488,1490,1492,1494,1496,1498,1500,1502,1504,1506,1508,1510,1512,1514,1516,1518,1520,1522,1524,1526,1528,1530,1532,1534,1536,1538,1540,1542,1544,1546,1548,1550,1552,1554,1556,1558,1560,1562,1564,1566,1568,1570,1572,1574,1576,1578,1580,1582,1584,1586,1588,1590,1592,1594,1596,1598,1600,1602,1604,1606,1608,1610,1612,1614,1616,1618,1620,1622,1624,1626,1628,1630,1632,1634,1636,1638,1640,1642,1644,1646,1648,1650,1652,1654,1656,1658,1660,1662,1664,1666,1668,1670,1672,1674,1676,1678,1680,1682,1684,1686,1688,1690,1692,1694,1696,1698,1700,1702,1704,1706,1708,1710,1712,1714,1716,1718,1720,1722,1724,1726,1728,1730,1732,1734,1736,1738,1740,1742,1744,1746,1748,1750,1752,1754,1756,1758,1760,1762,1764,1766,1768,1770,1772,1774,1776,1778,1780,1782,1784,1786,1788,1790,1792,1794,1796,1798,1800,1802,1804,1806,1808,1810,1812,1814,1816,1818,1820,1822,1824,1826,1828,1830,1832,1834,1836,1838,1840,1842,1844,1846,1848,1850,1852,1854,1856,1858,1860,1862,1864,1866,1868,1870,1872,1874,1876,1878,1880,1882,1884,1886,1888,1890,1892,1894,1896,1898,1900,1902,1904,1906,1908,1910,1912,1914,1916,1918,1920,1922,1924,1926,1928,1930,1932,1934,1936,1938,1940,1942,1944,1946,1948,1950,1952,1954,1956,1958,1960,1962,1964,1966,1968,1970,1972,1974,1976,1978,1980,1982,1984,1986,1988,1990,1992,1994,1996,1998,2000,2002,2004,2006,2008,2010,2012,2014,2016,2018,2020,2022,2024,2026,2028,2030,2032,2034,2036,2038,2040,2042,2044,2046,2048,2050,2052,2054,2056,2058,2060,2062,2064,2066,2068,2070,2072,2074,2076,2078,2080,2082,2084,2086,2088,2090,2092,2094,2096,2098,2100,2102,2104,2106,2108,2110,2112,2114,2116,2118,2120,2122,2124,2126,2128,2130,2132,2134,2136,2138,2140,2142,2144,2146,2148,2150,2152,2154,2156,2158,2160,2162,2164,2166,2168,2170,2172,2174,2176,2178,2180,2182,2184,2186,2188,2190,2192,2194,2196,2198,2200,2202,2204,2206,2208,2210,2212,2214,2216,2218,2220,2222,2224,2226,2228,2230,2232,2234,2236,2238,2240,2242,2244,2246,2248,2250,2252,2254,2256,2258,2260,2262,2264,2266,2268,2270,2272,2274,2276,2278,2280,2282,2284,2286,2288,2290,2292,2294,2296,2298,2300,2302,2304,2306,2308,2310,2312,2314,2316,2318,2320,2322,2324,2326,2328,2330,2332,2334,2336,2338,2340,2342,2344,2346,2348,2350,2352,2354,2356,2358,2360,2362,2364,2366,2368,2370,2372,2374,2376,2378,2380,2382,2384,2386,2388,2390,2392,2394,2396,2398,2400,2402,2404,2406,2408,2410,2412,2414,2416,2418,2420,2422,2424,2426,2428,2430,2432,2434,2436,2438,2440,2442,2444,2446,2448,2450,2452,2454,2456,2458,2460,2462,2464,2466,2468,2470,2472,2474,2476,2478,2480,2482,2484,2486,2488,2490,2492,2494,2496,2498,2500,2502,2504,2506,2508,2510,2512,2514,2516,2518,2520,2522,2524,2526,2528,2530,2532,2534,2536,2538,2540,2542,2544,2546,2548,2550,2552,2554,2556,2558,2560,2562,2564,2566,2568,2570,2572,2574,2576,2578,2580,2582,2584,2586,2588,2590,2592,2594,2596,2598,2600,2602,2604,2606,2608,2610,2612,2614,2616,2618,2620,2622,2624,2626,2628,2630,2632,2634,2636,2638,2640,2642,2644,2646,2648,2650,2652,2654,2656,2658,2660,2662,2664,2666,2668,2670,2672,2674,2676,2678,2680,2682,2684,2686,2688,2690,2692,2694,2696,2698,2700,2702,2704,2706,2708,2710,2712,2714,2716,2718,2720,2722,2724,2726,2728,2730,2732,2734,2736,2738,2740,2742,2744,2746,2748,2750,2752,2754,2756,2758,2760,2762,2764,2766,2768,2770,2772,2774,2776,2778,2780,2782,2784,2786,2788,2790,2792,2794,2796,2798,2800,2802,2804,2806,2808,2810,2812,2814,2816,2818,2820,2822,2824,2826,2828,2830,2832,2834,2836,2838,2840,2842,2844,2846,2848,2850,2852,2854,2856,2858,2860,2862,2864,2866,2868,2870,2872,2874,2876,2878,2880,2882,2884,2886,2888,2890,2892,2894,2896,2898,2900,2902,2904,2906,2908,2910,2912,2914,2916,2918,2920,2922,2924,2926,2928,2930,2932,2934,2936,2938,2940,2942,2944,2946,2948,2950,2952,2954,2956,2958,2960,2962,2964,2966,2968,2970,2972,2974,2976,2978,2980,2982,2984,2986,2988,2990,2992,2994,2996,2998,3000,3002,3004,3006,3008,3010,3012,3014,3016,3018,3020,3022,3024,3026,3028,3030,3032,3034,3036,3038,3040,3042,3044,3046,3048,3050,3052,3054,3056,3058,3060,3062,3064,3066,3068,3070,3072,3074,3076,3078,3080,3082,3084,3086,3088,3090,3092,3094,3096,3098,3100,3102,3104,3106,3108,3110,3112,3114,3116,3118,3120,3122,3124,3126,3128,3130,3132,3134,3136,3138,3140,3142,3144,3146,3148,3150,3152,3154,3156,3158,3160,3162,3164,3166,3168,3170,3172,3174,3176,3178,3180,3182,3184,3186,3188,3190,3192,3194,3196,3198,3200,3202,3204,3206,3208,3210,3212,3214,3216,3218,3220,3222,3224,3226,3228,3230,3232,3234,3236,3238,3240,3242,3244,3246,3248,3250,3252,3254,3256,3258,3260,3262,3264,3266,3268,3270,3272,3274,3276,3278,3280,3282,3284,3286,3288,3290,3292,3294,3296,3298,3300,3302,3304,3306,3308,3310,3312,3314,3316,3318,3320,3322,3324,3326,3328,3330,3332,3334,3336,3338,3340,3342,3344,3346,3348,3350,3352,3354,3356,3358,3360,3362,3364,3366,3368,3370,3372,3374,3376,3378,3380,3382,3384,3386,3388,3390,3392,3394,3396,3398,3400,3402,3404,3406,3408,3410,3412,3414,3416,3418,3420,3422,3424,3426,3428,3430,3432,3434,3436,3438,3440,3442,3444,3446,3448,3450,3452,3454,3456,3458,3460,3462,3464,3466,3468,3470,3472,3474,3476,3478,3480,3482,3484,3486,3488,3490,3492,3494,3496,3498,3500,3502,3504,3506,3508,3510,3512,3514,3516,3518,3520,3522,3524,3526,3528,3530,3532,3534,3536,3538,3540,3542,3544,3546,3548,3550,3552,3554,3556,3558,3560,3562,3564,3566,3568,3570,3572,3574,3576,3578,3580,3582,3584,3586,3588,3590,3592,3594,3596,3598,3600,3602,3604,3606,3608,3610,3612,3614,3616,3618,3620,3622,3624,3626,3628,3630,3632,3634,3636,3638,3640,3642,3644,3646,3648,3650,3652,3654,3656,3658,3660,3662,3664,3666,3668,3670,3672,3674,3676,3678,3680,3682,3684,3686,3688,3690,3692,3694,3696,3698,3700,3702,3704,3706,3708,3710,3712,3714,3716,3718,3720,3722,3724,3726,3728,3730,3732,3734,3736,3738,3740,3742,3744,3746,3748,3750,3752,3754,3756,3758,3760,3762,3764,3766,3768,3770,3772,3774,3776,3778,3780,3782,3784,3786,3788,3790,3792,3794,3796,3798,3800,3802,3804,3806,3808,3810,3812,3814,3816,3818,3820,3822,3824,3826,3828,3830,3832,3834,3836,3838,3840,3842,3844,3846,3848,3850,3852,3854,3856,3858,3860,3862,3864,3866,3868,3870,3872,3874,3876,3878,3880,3882,3884,3886,3888,3890,3892,3894,3896,3898,3900,3902,3904,3906,3908,3910,3912,3914,3916,3918,3920,3922,3924,3926,3928,3930,3932,3934,3936,3938,3940,3942,3944,3946,3948,3950,3952,3954,3956,3958,3960,3962,3964,3966,3968,3970,3972,3974,3976,3978,3980,3982,3984,3986,3988,3990,3992,3994,3996,3998,4000,4002,4004,4006,4008,4010,4012,4014,4016,4018,4020,4022,4024,4026,4028,4030,4032,4034,4036,4038,4040,4042,4044,4046,4048,4050,4052,4054,4056,4058,4060,4062,4064,4066,4068,4070,4072,4074,4076,4078,4080,4082,4084,4086,4088,4090,4092,4094,4096,4098,4100,4102,4104,4106,4108,4110,4112,4114,4116,4118,4120,4122,4124,4126,4128,4130,4132,4134,4136,4138,4140],{"categories":127},[128],"Developer Productivity",{"categories":130},[131],"Business & SaaS",{"categories":133},[90],{"categories":135},[136],"AI Automation",{"categories":138},[139],"Product Strategy",{"categories":141},[90],{"categories":143},[128],{"categories":145},[131],{"categories":147},[],{"categories":149},[90],{"categories":151},[],{"categories":153},[154],"AI News & Trends",{"categories":156},[136],{"categories":158},[136],{"categories":160},[154],{"categories":162},[136],{"categories":164},[136],{"categories":166},[136],{"categories":168},[90],{"categories":170},[90],{"categories":172},[90],{"categories":174},[154],{"categories":176},[90],{"categories":178},[90],{"categories":180},[],{"categories":182},[183],"Design & Frontend",{"categories":185},[186],"Data Science & Visualization",{"categories":188},[154],{"categories":190},[],{"categories":192},[193],"Software Engineering",{"categories":195},[90],{"categories":197},[136],{"categories":199},[200],"Marketing & Growth",{"categories":202},[183],{"categories":204},[90],{"categories":206},[136],{"categories":208},[],{"categories":210},[],{"categories":212},[183],{"categories":214},[136],{"categories":216},[128],{"categories":218},[193],{"categories":220},[183],{"categories":222},[90],{"categories":224},[225],"DevOps & Cloud",{"categories":227},[136],{"categories":229},[154],{"categories":231},[],{"categories":233},[],{"categories":235},[136],{"categories":237},[193],{"categories":239},[],{"categories":241},[131],{"categories":243},[],{"categories":245},[],{"categories":247},[136],{"categories":249},[90],{"categories":251},[90],{"categories":253},[136],{"categories":255},[90],{"categories":257},[90],{"categories":259},[],{"categories":261},[193],{"categories":263},[],{"categories":265},[],{"categories":267},[193],{"categories":269},[],{"categories":271},[193],{"categories":273},[90],{"categories":275},[90],{"categories":277},[200],{"categories":279},[183],{"categories":281},[183],{"categories":283},[90],{"categories":285},[136],{"categories":287},[193],{"categories":289},[90],{"categories":291},[90],{"categories":293},[136],{"categories":295},[136],{"categories":297},[186],{"categories":299},[154],{"categories":301},[136],{"categories":303},[136],{"categories":305},[200],{"categories":307},[136],{"categories":309},[139],{"categories":311},[193],{"categories":313},[],{"categories":315},[136],{"categories":317},[],{"categories":319},[136],{"categories":321},[193],{"categories":323},[225],{"categories":325},[183],{"categories":327},[90],{"categories":329},[],{"categories":331},[90],{"categories":333},[],{"categories":335},[136],{"categories":337},[],{"categories":339},[90],{"categories":341},[],{"categories":343},[128],{"categories":345},[193],{"categories":347},[131],{"categories":349},[90],{"categories":351},[90],{"categories":353},[154],{"categories":355},[90],{"categories":357},[],{"categories":359},[90],{"categories":361},[],{"categories":363},[193],{"categories":365},[186],{"categories":367},[],{"categories":369},[90],{"categories":371},[183],{"categories":373},[],{"categories":375},[183],{"categories":377},[136],{"categories":379},[],{"categories":381},[90],{"categories":383},[136],{"categories":385},[154],{"categories":387},[131],{"categories":389},[90],{"categories":391},[],{"categories":393},[136],{"categories":395},[90],{"categories":397},[139],{"categories":399},[],{"categories":401},[90],{"categories":403},[139],{"categories":405},[136],{"categories":407},[136],{"categories":409},[],{"categories":411},[186],{"categories":413},[90],{"categories":415},[],{"categories":417},[128],{"categories":419},[131],{"categories":421},[90],{"categories":423},[136],{"categories":425},[193],{"categories":427},[90],{"categories":429},[],{"categories":431},[],{"categories":433},[90],{"categories":435},[90],{"categories":437},[],{"categories":439},[183],{"categories":441},[],{"categories":443},[90],{"categories":445},[],{"categories":447},[136],{"categories":449},[90],{"categories":451},[183],{"categories":453},[],{"categories":455},[90],{"categories":457},[90],{"categories":459},[131],{"categories":461},[136],{"categories":463},[90],{"categories":465},[90],{"categories":467},[183],{"categories":469},[136],{"categories":471},[],{"categories":473},[],{"categories":475},[154],{"categories":477},[],{"categories":479},[90],{"categories":481},[131,200],{"categories":483},[],{"categories":485},[90],{"categories":487},[136],{"categories":489},[],{"categories":491},[],{"categories":493},[90],{"categories":495},[],{"categories":497},[90],{"categories":499},[225],{"categories":501},[],{"categories":503},[154],{"categories":505},[183],{"categories":507},[],{"categories":509},[154],{"categories":511},[136],{"categories":513},[154],{"categories":515},[90],{"categories":517},[200],{"categories":519},[],{"categories":521},[131],{"categories":523},[90],{"categories":525},[136],{"categories":527},[],{"categories":529},[90,225],{"categories":531},[90],{"categories":533},[90],{"categories":535},[90],{"categories":537},[136],{"categories":539},[90,193],{"categories":541},[186],{"categories":543},[90],{"categories":545},[200],{"categories":547},[136],{"categories":549},[90],{"categories":551},[136],{"categories":553},[],{"categories":555},[136],{"categories":557},[90],{"categories":559},[90,131],{"categories":561},[],{"categories":563},[183],{"categories":565},[183],{"categories":567},[],{"categories":569},[],{"categories":571},[154],{"categories":573},[],{"categories":575},[128],{"categories":577},[193],{"categories":579},[90],{"categories":581},[183],{"categories":583},[136],{"categories":585},[193],{"categories":587},[154],{"categories":589},[183],{"categories":591},[],{"categories":593},[90],{"categories":595},[90],{"categories":597},[90],{"categories":599},[90],{"categories":601},[154],{"categories":603},[128],{"categories":605},[90],{"categories":607},[136],{"categories":609},[225],{"categories":611},[183],{"categories":613},[136],{"categories":615},[],{"categories":617},[],{"categories":619},[183],{"categories":621},[154],{"categories":623},[186],{"categories":625},[],{"categories":627},[90],{"categories":629},[90],{"categories":631},[131],{"categories":633},[90],{"categories":635},[90],{"categories":637},[90],{"categories":639},[154],{"categories":641},[],{"categories":643},[136],{"categories":645},[193],{"categories":647},[],{"categories":649},[90],{"categories":651},[90],{"categories":653},[136],{"categories":655},[],{"categories":657},[],{"categories":659},[90],{"categories":661},[],{"categories":663},[131],{"categories":665},[136],{"categories":667},[136],{"categories":669},[],{"categories":671},[128],{"categories":673},[90],{"categories":675},[131],{"categories":677},[154],{"categories":679},[128],{"categories":681},[],{"categories":683},[],{"categories":685},[],{"categories":687},[154],{"categories":689},[154],{"categories":691},[],{"categories":693},[193],{"categories":695},[],{"categories":697},[131],{"categories":699},[],{"categories":701},[],{"categories":703},[128],{"categories":705},[],{"categories":707},[200],{"categories":709},[136],{"categories":711},[131],{"categories":713},[136],{"categories":715},[193],{"categories":717},[],{"categories":719},[139],{"categories":721},[183],{"categories":723},[193],{"categories":725},[90],{"categories":727},[136],{"categories":729},[131],{"categories":731},[90],{"categories":733},[],{"categories":735},[],{"categories":737},[193],{"categories":739},[186],{"categories":741},[139],{"categories":743},[136],{"categories":745},[90],{"categories":747},[],{"categories":749},[225],{"categories":751},[],{"categories":753},[136],{"categories":755},[],{"categories":757},[128],{"categories":759},[],{"categories":761},[90],{"categories":763},[90],{"categories":765},[183],{"categories":767},[200],{"categories":769},[136],{"categories":771},[],{"categories":773},[193],{"categories":775},[128],{"categories":777},[],{"categories":779},[154],{"categories":781},[90,225],{"categories":783},[90],{"categories":785},[154],{"categories":787},[90],{"categories":789},[90],{"categories":791},[131],{"categories":793},[90],{"categories":795},[],{"categories":797},[90],{"categories":799},[131],{"categories":801},[],{"categories":803},[136],{"categories":805},[193],{"categories":807},[183],{"categories":809},[154],{"categories":811},[186],{"categories":813},[128],{"categories":815},[90],{"categories":817},[136],{"categories":819},[193],{"categories":821},[],{"categories":823},[],{"categories":825},[139],{"categories":827},[],{"categories":829},[90],{"categories":831},[],{"categories":833},[183],{"categories":835},[193],{"categories":837},[183],{"categories":839},[90],{"categories":841},[183],{"categories":843},[],{"categories":845},[],{"categories":847},[154],{"categories":849},[136],{"categories":851},[136],{"categories":853},[90],{"categories":855},[90],{"categories":857},[90],{"categories":859},[131],{"categories":861},[90],{"categories":863},[],{"categories":865},[193],{"categories":867},[193],{"categories":869},[131],{"categories":871},[],{"categories":873},[90],{"categories":875},[90],{"categories":877},[131],{"categories":879},[154],{"categories":881},[200],{"categories":883},[90],{"categories":885},[136],{"categories":887},[],{"categories":889},[183],{"categories":891},[],{"categories":893},[90],{"categories":895},[90],{"categories":897},[],{"categories":899},[131],{"categories":901},[136],{"categories":903},[],{"categories":905},[225],{"categories":907},[186],{"categories":909},[193],{"categories":911},[200],{"categories":913},[90],{"categories":915},[193],{"categories":917},[136],{"categories":919},[],{"categories":921},[],{"categories":923},[136],{"categories":925},[128],{"categories":927},[136],{"categories":929},[139],{"categories":931},[131],{"categories":933},[],{"categories":935},[90],{"categories":937},[139],{"categories":939},[90],{"categories":941},[90],{"categories":943},[200],{"categories":945},[90],{"categories":947},[183],{"categories":949},[136],{"categories":951},[],{"categories":953},[],{"categories":955},[225],{"categories":957},[193],{"categories":959},[],{"categories":961},[136],{"categories":963},[90],{"categories":965},[183,90],{"categories":967},[128],{"categories":969},[],{"categories":971},[90],{"categories":973},[128],{"categories":975},[183],{"categories":977},[136],{"categories":979},[193],{"categories":981},[],{"categories":983},[90],{"categories":985},[],{"categories":987},[],{"categories":989},[90],{"categories":991},[128],{"categories":993},[90],{"categories":995},[],{"categories":997},[136],{"categories":999},[139],{"categories":1001},[90],{"categories":1003},[90],{"categories":1005},[90],{"categories":1007},[183],{"categories":1009},[136],{"categories":1011},[225],{"categories":1013},[183],{"categories":1015},[136],{"categories":1017},[90],{"categories":1019},[90],{"categories":1021},[90],{"categories":1023},[193],{"categories":1025},[],{"categories":1027},[154],{"categories":1029},[],{"categories":1031},[139],{"categories":1033},[136],{"categories":1035},[183],{"categories":1037},[90],{"categories":1039},[136],{"categories":1041},[193],{"categories":1043},[183],{"categories":1045},[136],{"categories":1047},[154],{"categories":1049},[],{"categories":1051},[90],{"categories":1053},[183],{"categories":1055},[90],{"categories":1057},[128],{"categories":1059},[154],{"categories":1061},[90],{"categories":1063},[200],{"categories":1065},[90],{"categories":1067},[136],{"categories":1069},[136],{"categories":1071},[90],{"categories":1073},[136],{"categories":1075},[136],{"categories":1077},[90],{"categories":1079},[136],{"categories":1081},[183],{"categories":1083},[90],{"categories":1085},[],{"categories":1087},[],{"categories":1089},[193],{"categories":1091},[],{"categories":1093},[128],{"categories":1095},[225],{"categories":1097},[90],{"categories":1099},[],{"categories":1101},[128],{"categories":1103},[131],{"categories":1105},[200],{"categories":1107},[],{"categories":1109},[131],{"categories":1111},[],{"categories":1113},[90],{"categories":1115},[193],{"categories":1117},[],{"categories":1119},[],{"categories":1121},[],{"categories":1123},[],{"categories":1125},[90],{"categories":1127},[136],{"categories":1129},[225],{"categories":1131},[128],{"categories":1133},[193],{"categories":1135},[90],{"categories":1137},[193],{"categories":1139},[139],{"categories":1141},[90],{"categories":1143},[200],{"categories":1145},[131],{"categories":1147},[90],{"categories":1149},[90],{"categories":1151},[90],{"categories":1153},[90,128],{"categories":1155},[193],{"categories":1157},[193],{"categories":1159},[183],{"categories":1161},[136],{"categories":1163},[90],{"categories":1165},[],{"categories":1167},[],{"categories":1169},[],{"categories":1171},[193],{"categories":1173},[186],{"categories":1175},[154],{"categories":1177},[183],{"categories":1179},[193],{"categories":1181},[],{"categories":1183},[90],{"categories":1185},[90],{"categories":1187},[],{"categories":1189},[136],{"categories":1191},[90],{"categories":1193},[90],{"categories":1195},[],{"categories":1197},[136],{"categories":1199},[90],{"categories":1201},[131],{"categories":1203},[],{"categories":1205},[128],{"categories":1207},[90],{"categories":1209},[128],{"categories":1211},[90],{"categories":1213},[193],{"categories":1215},[200],{"categories":1217},[136],{"categories":1219},[90,183],{"categories":1221},[154],{"categories":1223},[90],{"categories":1225},[183],{"categories":1227},[],{"categories":1229},[193],{"categories":1231},[225],{"categories":1233},[183],{"categories":1235},[136],{"categories":1237},[],{"categories":1239},[],{"categories":1241},[],{"categories":1243},[],{"categories":1245},[193],{"categories":1247},[136],{"categories":1249},[136],{"categories":1251},[225],{"categories":1253},[90],{"categories":1255},[90],{"categories":1257},[136],{"categories":1259},[90],{"categories":1261},[90],{"categories":1263},[],{"categories":1265},[183],{"categories":1267},[],{"categories":1269},[],{"categories":1271},[136],{"categories":1273},[],{"categories":1275},[],{"categories":1277},[200],{"categories":1279},[200],{"categories":1281},[136],{"categories":1283},[193],{"categories":1285},[],{"categories":1287},[90],{"categories":1289},[90],{"categories":1291},[193],{"categories":1293},[183],{"categories":1295},[183],{"categories":1297},[136],{"categories":1299},[128],{"categories":1301},[90],{"categories":1303},[183],{"categories":1305},[183],{"categories":1307},[136],{"categories":1309},[136],{"categories":1311},[90],{"categories":1313},[],{"categories":1315},[90],{"categories":1317},[],{"categories":1319},[90],{"categories":1321},[136],{"categories":1323},[154],{"categories":1325},[193],{"categories":1327},[90],{"categories":1329},[128],{"categories":1331},[90],{"categories":1333},[],{"categories":1335},[136],{"categories":1337},[136],{"categories":1339},[],{"categories":1341},[90],{"categories":1343},[128],{"categories":1345},[90],{"categories":1347},[128],{"categories":1349},[128],{"categories":1351},[],{"categories":1353},[],{"categories":1355},[136],{"categories":1357},[154],{"categories":1359},[136],{"categories":1361},[90],{"categories":1363},[90],{"categories":1365},[154],{"categories":1367},[186],{"categories":1369},[139],{"categories":1371},[154],{"categories":1373},[183],{"categories":1375},[],{"categories":1377},[],{"categories":1379},[154],{"categories":1381},[],{"categories":1383},[],{"categories":1385},[],{"categories":1387},[],{"categories":1389},[193],{"categories":1391},[193],{"categories":1393},[186],{"categories":1395},[],{"categories":1397},[90],{"categories":1399},[90],{"categories":1401},[186],{"categories":1403},[193],{"categories":1405},[],{"categories":1407},[],{"categories":1409},[136],{"categories":1411},[136],{"categories":1413},[154],{"categories":1415},[154],{"categories":1417},[136],{"categories":1419},[136],{"categories":1421},[128],{"categories":1423},[90,225],{"categories":1425},[],{"categories":1427},[183],{"categories":1429},[128],{"categories":1431},[136],{"categories":1433},[183],{"categories":1435},[],{"categories":1437},[136],{"categories":1439},[136],{"categories":1441},[90],{"categories":1443},[200],{"categories":1445},[193],{"categories":1447},[183],{"categories":1449},[],{"categories":1451},[136],{"categories":1453},[90],{"categories":1455},[136],{"categories":1457},[136],{"categories":1459},[136],{"categories":1461},[200],{"categories":1463},[90],{"categories":1465},[136],{"categories":1467},[90],{"categories":1469},[],{"categories":1471},[200],{"categories":1473},[154],{"categories":1475},[193],{"categories":1477},[90],{"categories":1479},[136],{"categories":1481},[],{"categories":1483},[],{"categories":1485},[90],{"categories":1487},[136],{"categories":1489},[154],{"categories":1491},[136],{"categories":1493},[136],{"categories":1495},[],{"categories":1497},[90],{"categories":1499},[],{"categories":1501},[],{"categories":1503},[136],{"categories":1505},[],{"categories":1507},[],{"categories":1509},[186],{"categories":1511},[90],{"categories":1513},[186],{"categories":1515},[154],{"categories":1517},[90],{"categories":1519},[90],{"categories":1521},[136],{"categories":1523},[90],{"categories":1525},[],{"categories":1527},[],{"categories":1529},[225],{"categories":1531},[90],{"categories":1533},[],{"categories":1535},[],{"categories":1537},[128],{"categories":1539},[],{"categories":1541},[],{"categories":1543},[90],{"categories":1545},[],{"categories":1547},[],{"categories":1549},[193],{"categories":1551},[154],{"categories":1553},[200],{"categories":1555},[131],{"categories":1557},[90],{"categories":1559},[90],{"categories":1561},[131],{"categories":1563},[],{"categories":1565},[183],{"categories":1567},[136],{"categories":1569},[131],{"categories":1571},[90],{"categories":1573},[90],{"categories":1575},[128],{"categories":1577},[90],{"categories":1579},[],{"categories":1581},[128],{"categories":1583},[90],{"categories":1585},[200],{"categories":1587},[136],{"categories":1589},[154],{"categories":1591},[131],{"categories":1593},[90],{"categories":1595},[90],{"categories":1597},[136],{"categories":1599},[],{"categories":1601},[90],{"categories":1603},[128],{"categories":1605},[90],{"categories":1607},[90],{"categories":1609},[],{"categories":1611},[154],{"categories":1613},[90],{"categories":1615},[90],{"categories":1617},[],{"categories":1619},[131],{"categories":1621},[131],{"categories":1623},[90],{"categories":1625},[90],{"categories":1627},[],{"categories":1629},[],{"categories":1631},[],{"categories":1633},[90],{"categories":1635},[154],{"categories":1637},[],{"categories":1639},[225],{"categories":1641},[90],{"categories":1643},[90],{"categories":1645},[],{"categories":1647},[90],{"categories":1649},[90],{"categories":1651},[90],{"categories":1653},[90,225],{"categories":1655},[90],{"categories":1657},[90],{"categories":1659},[183],{"categories":1661},[136],{"categories":1663},[],{"categories":1665},[136],{"categories":1667},[136],{"categories":1669},[90],{"categories":1671},[90],{"categories":1673},[90],{"categories":1675},[128],{"categories":1677},[128],{"categories":1679},[193],{"categories":1681},[183],{"categories":1683},[136],{"categories":1685},[],{"categories":1687},[90],{"categories":1689},[154],{"categories":1691},[90],{"categories":1693},[90],{"categories":1695},[131],{"categories":1697},[],{"categories":1699},[225],{"categories":1701},[183],{"categories":1703},[183],{"categories":1705},[136],{"categories":1707},[154],{"categories":1709},[136],{"categories":1711},[90],{"categories":1713},[],{"categories":1715},[90],{"categories":1717},[],{"categories":1719},[],{"categories":1721},[90],{"categories":1723},[90],{"categories":1725},[90],{"categories":1727},[136],{"categories":1729},[90],{"categories":1731},[90],{"categories":1733},[],{"categories":1735},[186],{"categories":1737},[136],{"categories":1739},[],{"categories":1741},[],{"categories":1743},[90],{"categories":1745},[90],{"categories":1747},[90],{"categories":1749},[154],{"categories":1751},[],{"categories":1753},[183],{"categories":1755},[225],{"categories":1757},[154],{"categories":1759},[193],{"categories":1761},[193],{"categories":1763},[154],{"categories":1765},[154],{"categories":1767},[225],{"categories":1769},[],{"categories":1771},[154],{"categories":1773},[90],{"categories":1775},[128],{"categories":1777},[90],{"categories":1779},[154],{"categories":1781},[],{"categories":1783},[193],{"categories":1785},[186],{"categories":1787},[90],{"categories":1789},[154],{"categories":1791},[193],{"categories":1793},[136],{"categories":1795},[154],{"categories":1797},[225],{"categories":1799},[136],{"categories":1801},[90],{"categories":1803},[90],{"categories":1805},[90],{"categories":1807},[],{"categories":1809},[131],{"categories":1811},[],{"categories":1813},[],{"categories":1815},[90],{"categories":1817},[90],{"categories":1819},[90],{"categories":1821},[90],{"categories":1823},[],{"categories":1825},[186],{"categories":1827},[128],{"categories":1829},[],{"categories":1831},[90],{"categories":1833},[90],{"categories":1835},[225],{"categories":1837},[225],{"categories":1839},[],{"categories":1841},[136],{"categories":1843},[154],{"categories":1845},[154],{"categories":1847},[90],{"categories":1849},[136],{"categories":1851},[],{"categories":1853},[183],{"categories":1855},[90],{"categories":1857},[90],{"categories":1859},[],{"categories":1861},[90],{"categories":1863},[],{"categories":1865},[193],{"categories":1867},[225],{"categories":1869},[90],{"categories":1871},[193],{"categories":1873},[131],{"categories":1875},[90],{"categories":1877},[],{"categories":1879},[136],{"categories":1881},[128],{"categories":1883},[128],{"categories":1885},[],{"categories":1887},[90],{"categories":1889},[183],{"categories":1891},[136],{"categories":1893},[],{"categories":1895},[90],{"categories":1897},[90],{"categories":1899},[136],{"categories":1901},[],{"categories":1903},[136],{"categories":1905},[193],{"categories":1907},[],{"categories":1909},[90],{"categories":1911},[],{"categories":1913},[90],{"categories":1915},[],{"categories":1917},[90],{"categories":1919},[90],{"categories":1921},[],{"categories":1923},[90],{"categories":1925},[154],{"categories":1927},[90],{"categories":1929},[90],{"categories":1931},[128],{"categories":1933},[90],{"categories":1935},[154],{"categories":1937},[136],{"categories":1939},[],{"categories":1941},[90],{"categories":1943},[183],{"categories":1945},[200],{"categories":1947},[90],{"categories":1949},[],{"categories":1951},[],{"categories":1953},[],{"categories":1955},[128],{"categories":1957},[154],{"categories":1959},[136],{"categories":1961},[90],{"categories":1963},[183],{"categories":1965},[136],{"categories":1967},[],{"categories":1969},[136],{"categories":1971},[],{"categories":1973},[90],{"categories":1975},[136],{"categories":1977},[90],{"categories":1979},[],{"categories":1981},[90],{"categories":1983},[90],{"categories":1985},[154],{"categories":1987},[183],{"categories":1989},[136],{"categories":1991},[183],{"categories":1993},[131],{"categories":1995},[],{"categories":1997},[],{"categories":1999},[90],{"categories":2001},[128],{"categories":2003},[154],{"categories":2005},[],{"categories":2007},[183],{"categories":2009},[],{"categories":2011},[193],{"categories":2013},[193],{"categories":2015},[183],{"categories":2017},[],{"categories":2019},[90],{"categories":2021},[],{"categories":2023},[200],{"categories":2025},[90],{"categories":2027},[225],{"categories":2029},[193],{"categories":2031},[],{"categories":2033},[136],{"categories":2035},[90],{"categories":2037},[128],{"categories":2039},[136],{"categories":2041},[136],{"categories":2043},[90],{"categories":2045},[],{"categories":2047},[128],{"categories":2049},[90],{"categories":2051},[131],{"categories":2053},[193],{"categories":2055},[183],{"categories":2057},[],{"categories":2059},[],{"categories":2061},[],{"categories":2063},[136],{"categories":2065},[193],{"categories":2067},[183],{"categories":2069},[154],{"categories":2071},[90],{"categories":2073},[154],{"categories":2075},[183],{"categories":2077},[],{"categories":2079},[183],{"categories":2081},[154],{"categories":2083},[131],{"categories":2085},[193],{"categories":2087},[90],{"categories":2089},[154],{"categories":2091},[200],{"categories":2093},[],{"categories":2095},[],{"categories":2097},[186],{"categories":2099},[90,193],{"categories":2101},[154],{"categories":2103},[90],{"categories":2105},[136],{"categories":2107},[90],{"categories":2109},[136],{"categories":2111},[90],{"categories":2113},[90],{"categories":2115},[],{"categories":2117},[193],{"categories":2119},[90],{"categories":2121},[186],{"categories":2123},[136],{"categories":2125},[200],{"categories":2127},[225],{"categories":2129},[],{"categories":2131},[128],{"categories":2133},[136],{"categories":2135},[136],{"categories":2137},[193],{"categories":2139},[90],{"categories":2141},[90],{"categories":2143},[],{"categories":2145},[],{"categories":2147},[],{"categories":2149},[225],{"categories":2151},[154],{"categories":2153},[90],{"categories":2155},[90],{"categories":2157},[90],{"categories":2159},[],{"categories":2161},[186],{"categories":2163},[131],{"categories":2165},[],{"categories":2167},[90],{"categories":2169},[136],{"categories":2171},[225],{"categories":2173},[],{"categories":2175},[183],{"categories":2177},[183],{"categories":2179},[],{"categories":2181},[193],{"categories":2183},[90],{"categories":2185},[183],{"categories":2187},[90],{"categories":2189},[],{"categories":2191},[154],{"categories":2193},[90],{"categories":2195},[90],{"categories":2197},[183],{"categories":2199},[136],{"categories":2201},[154],{"categories":2203},[],{"categories":2205},[136],{"categories":2207},[183],{"categories":2209},[90],{"categories":2211},[],{"categories":2213},[90],{"categories":2215},[90],{"categories":2217},[225],{"categories":2219},[154],{"categories":2221},[186],{"categories":2223},[186],{"categories":2225},[],{"categories":2227},[],{"categories":2229},[],{"categories":2231},[136],{"categories":2233},[193],{"categories":2235},[193],{"categories":2237},[90],{"categories":2239},[90],{"categories":2241},[],{"categories":2243},[],{"categories":2245},[90],{"categories":2247},[],{"categories":2249},[136],{"categories":2251},[90],{"categories":2253},[],{"categories":2255},[90],{"categories":2257},[131],{"categories":2259},[90],{"categories":2261},[200],{"categories":2263},[136],{"categories":2265},[90],{"categories":2267},[90],{"categories":2269},[90],{"categories":2271},[193],{"categories":2273},[],{"categories":2275},[154],{"categories":2277},[136],{"categories":2279},[],{"categories":2281},[154],{"categories":2283},[136],{"categories":2285},[90],{"categories":2287},[136],{"categories":2289},[],{"categories":2291},[131],{"categories":2293},[136],{"categories":2295},[],{"categories":2297},[193],{"categories":2299},[90],{"categories":2301},[128],{"categories":2303},[154],{"categories":2305},[225],{"categories":2307},[136],{"categories":2309},[136],{"categories":2311},[128],{"categories":2313},[],{"categories":2315},[90],{"categories":2317},[],{"categories":2319},[],{"categories":2321},[183],{"categories":2323},[90,131],{"categories":2325},[90],{"categories":2327},[],{"categories":2329},[128],{"categories":2331},[186],{"categories":2333},[90],{"categories":2335},[193],{"categories":2337},[90],{"categories":2339},[136],{"categories":2341},[90],{"categories":2343},[90],{"categories":2345},[90],{"categories":2347},[154],{"categories":2349},[136],{"categories":2351},[90],{"categories":2353},[],{"categories":2355},[],{"categories":2357},[136],{"categories":2359},[90],{"categories":2361},[225],{"categories":2363},[],{"categories":2365},[90],{"categories":2367},[136],{"categories":2369},[],{"categories":2371},[136],{"categories":2373},[90],{"categories":2375},[200],{"categories":2377},[186],{"categories":2379},[136],{"categories":2381},[90],{"categories":2383},[225],{"categories":2385},[],{"categories":2387},[90],{"categories":2389},[200],{"categories":2391},[183],{"categories":2393},[90],{"categories":2395},[90],{"categories":2397},[],{"categories":2399},[200],{"categories":2401},[154],{"categories":2403},[90],{"categories":2405},[90],{"categories":2407},[128],{"categories":2409},[],{"categories":2411},[],{"categories":2413},[183],{"categories":2415},[90],{"categories":2417},[186],{"categories":2419},[200],{"categories":2421},[136],{"categories":2423},[200],{"categories":2425},[154],{"categories":2427},[],{"categories":2429},[],{"categories":2431},[90],{"categories":2433},[136],{"categories":2435},[90],{"categories":2437},[90],{"categories":2439},[],{"categories":2441},[90,193],{"categories":2443},[154],{"categories":2445},[136],{"categories":2447},[193],{"categories":2449},[90],{"categories":2451},[128],{"categories":2453},[],{"categories":2455},[],{"categories":2457},[128],{"categories":2459},[193],{"categories":2461},[200],{"categories":2463},[90],{"categories":2465},[193],{"categories":2467},[],{"categories":2469},[183,90],{"categories":2471},[225],{"categories":2473},[128],{"categories":2475},[],{"categories":2477},[131],{"categories":2479},[131],{"categories":2481},[90],{"categories":2483},[90],{"categories":2485},[193],{"categories":2487},[136],{"categories":2489},[154],{"categories":2491},[200],{"categories":2493},[183],{"categories":2495},[90],{"categories":2497},[90],{"categories":2499},[90],{"categories":2501},[128],{"categories":2503},[90],{"categories":2505},[136],{"categories":2507},[154],{"categories":2509},[],{"categories":2511},[],{"categories":2513},[186],{"categories":2515},[193],{"categories":2517},[90],{"categories":2519},[183],{"categories":2521},[90],{"categories":2523},[186],{"categories":2525},[90],{"categories":2527},[90],{"categories":2529},[90],{"categories":2531},[136],{"categories":2533},[136],{"categories":2535},[90,131],{"categories":2537},[],{"categories":2539},[183],{"categories":2541},[],{"categories":2543},[90],{"categories":2545},[154],{"categories":2547},[128],{"categories":2549},[128],{"categories":2551},[136],{"categories":2553},[90],{"categories":2555},[90],{"categories":2557},[131],{"categories":2559},[193],{"categories":2561},[200],{"categories":2563},[90],{"categories":2565},[],{"categories":2567},[154],{"categories":2569},[90],{"categories":2571},[90],{"categories":2573},[90],{"categories":2575},[90],{"categories":2577},[90],{"categories":2579},[193],{"categories":2581},[154],{"categories":2583},[193],{"categories":2585},[193],{"categories":2587},[90],{"categories":2589},[90],{"categories":2591},[136],{"categories":2593},[154],{"categories":2595},[90],{"categories":2597},[183],{"categories":2599},[90],{"categories":2601},[90],{"categories":2603},[225],{"categories":2605},[90],{"categories":2607},[139],{"categories":2609},[136],{"categories":2611},[90],{"categories":2613},[154],{"categories":2615},[136],{"categories":2617},[200],{"categories":2619},[90],{"categories":2621},[],{"categories":2623},[90],{"categories":2625},[90],{"categories":2627},[],{"categories":2629},[],{"categories":2631},[],{"categories":2633},[131],{"categories":2635},[90],{"categories":2637},[136],{"categories":2639},[154],{"categories":2641},[154],{"categories":2643},[154],{"categories":2645},[154],{"categories":2647},[],{"categories":2649},[128],{"categories":2651},[136],{"categories":2653},[154],{"categories":2655},[90],{"categories":2657},[128],{"categories":2659},[136],{"categories":2661},[90],{"categories":2663},[90,136],{"categories":2665},[136],{"categories":2667},[225],{"categories":2669},[154],{"categories":2671},[154],{"categories":2673},[136],{"categories":2675},[90],{"categories":2677},[],{"categories":2679},[154],{"categories":2681},[200],{"categories":2683},[128],{"categories":2685},[90],{"categories":2687},[90],{"categories":2689},[],{"categories":2691},[193],{"categories":2693},[],{"categories":2695},[128],{"categories":2697},[136],{"categories":2699},[154],{"categories":2701},[90],{"categories":2703},[154],{"categories":2705},[128],{"categories":2707},[154],{"categories":2709},[154],{"categories":2711},[],{"categories":2713},[131],{"categories":2715},[136],{"categories":2717},[154],{"categories":2719},[154],{"categories":2721},[154],{"categories":2723},[154],{"categories":2725},[154],{"categories":2727},[154],{"categories":2729},[154],{"categories":2731},[154],{"categories":2733},[154],{"categories":2735},[154],{"categories":2737},[186],{"categories":2739},[128],{"categories":2741},[90],{"categories":2743},[90],{"categories":2745},[136],{"categories":2747},[],{"categories":2749},[90,128],{"categories":2751},[],{"categories":2753},[136],{"categories":2755},[154],{"categories":2757},[136],{"categories":2759},[90],{"categories":2761},[90],{"categories":2763},[90],{"categories":2765},[90],{"categories":2767},[90],{"categories":2769},[136],{"categories":2771},[131],{"categories":2773},[],{"categories":2775},[183],{"categories":2777},[154],{"categories":2779},[90],{"categories":2781},[],{"categories":2783},[],{"categories":2785},[136],{"categories":2787},[183],{"categories":2789},[90],{"categories":2791},[],{"categories":2793},[90],{"categories":2795},[],{"categories":2797},[200],{"categories":2799},[90],{"categories":2801},[],{"categories":2803},[],{"categories":2805},[154],{"categories":2807},[128],{"categories":2809},[90],{"categories":2811},[131],{"categories":2813},[90],{"categories":2815},[131],{"categories":2817},[183],{"categories":2819},[],{"categories":2821},[154],{"categories":2823},[],{"categories":2825},[183],{"categories":2827},[90],{"categories":2829},[200],{"categories":2831},[],{"categories":2833},[200],{"categories":2835},[],{"categories":2837},[],{"categories":2839},[136],{"categories":2841},[],{"categories":2843},[131],{"categories":2845},[128],{"categories":2847},[183],{"categories":2849},[193],{"categories":2851},[],{"categories":2853},[],{"categories":2855},[90],{"categories":2857},[128],{"categories":2859},[200],{"categories":2861},[],{"categories":2863},[136],{"categories":2865},[136],{"categories":2867},[154],{"categories":2869},[193],{"categories":2871},[90],{"categories":2873},[136],{"categories":2875},[90],{"categories":2877},[136],{"categories":2879},[90],{"categories":2881},[139],{"categories":2883},[200],{"categories":2885},[154],{"categories":2887},[],{"categories":2889},[200],{"categories":2891},[],{"categories":2893},[193],{"categories":2895},[136],{"categories":2897},[],{"categories":2899},[90],{"categories":2901},[136],{"categories":2903},[131],{"categories":2905},[128],{"categories":2907},[90],{"categories":2909},[183],{"categories":2911},[193],{"categories":2913},[193],{"categories":2915},[90],{"categories":2917},[186],{"categories":2919},[90],{"categories":2921},[136],{"categories":2923},[131],{"categories":2925},[183],{"categories":2927},[136],{"categories":2929},[90],{"categories":2931},[90],{"categories":2933},[136],{"categories":2935},[154],{"categories":2937},[],{"categories":2939},[128],{"categories":2941},[90],{"categories":2943},[90],{"categories":2945},[136],{"categories":2947},[90],{"categories":2949},[90],{"categories":2951},[],{"categories":2953},[183],{"categories":2955},[131],{"categories":2957},[154],{"categories":2959},[90],{"categories":2961},[90],{"categories":2963},[183],{"categories":2965},[90],{"categories":2967},[200],{"categories":2969},[186],{"categories":2971},[90],{"categories":2973},[154],{"categories":2975},[90],{"categories":2977},[136],{"categories":2979},[225],{"categories":2981},[90],{"categories":2983},[136],{"categories":2985},[186],{"categories":2987},[],{"categories":2989},[136],{"categories":2991},[193],{"categories":2993},[183],{"categories":2995},[90],{"categories":2997},[128],{"categories":2999},[193],{"categories":3001},[131],{"categories":3003},[193],{"categories":3005},[90],{"categories":3007},[],{"categories":3009},[136],{"categories":3011},[136],{"categories":3013},[90],{"categories":3015},[186],{"categories":3017},[],{"categories":3019},[154],{"categories":3021},[],{"categories":3023},[154],{"categories":3025},[90],{"categories":3027},[90],{"categories":3029},[136],{"categories":3031},[136],{"categories":3033},[136],{"categories":3035},[],{"categories":3037},[154],{"categories":3039},[],{"categories":3041},[90],{"categories":3043},[90],{"categories":3045},[],{"categories":3047},[183],{"categories":3049},[136],{"categories":3051},[200],{"categories":3053},[128],{"categories":3055},[],{"categories":3057},[90],{"categories":3059},[],{"categories":3061},[128],{"categories":3063},[154],{"categories":3065},[193],{"categories":3067},[90],{"categories":3069},[90],{"categories":3071},[90],{"categories":3073},[193],{"categories":3075},[154],{"categories":3077},[183],{"categories":3079},[90],{"categories":3081},[90],{"categories":3083},[90],{"categories":3085},[154],{"categories":3087},[90],{"categories":3089},[154],{"categories":3091},[154],{"categories":3093},[136],{"categories":3095},[136],{"categories":3097},[193],{"categories":3099},[154],{"categories":3101},[136],{"categories":3103},[90],{"categories":3105},[193],{"categories":3107},[183],{"categories":3109},[],{"categories":3111},[136],{"categories":3113},[],{"categories":3115},[],{"categories":3117},[],{"categories":3119},[131],{"categories":3121},[136],{"categories":3123},[90],{"categories":3125},[136],{"categories":3127},[128],{"categories":3129},[136],{"categories":3131},[200],{"categories":3133},[],{"categories":3135},[136],{"categories":3137},[],{"categories":3139},[128],{"categories":3141},[136],{"categories":3143},[],{"categories":3145},[136],{"categories":3147},[90],{"categories":3149},[154],{"categories":3151},[90],{"categories":3153},[136],{"categories":3155},[154],{"categories":3157},[136],{"categories":3159},[193],{"categories":3161},[183],{"categories":3163},[128],{"categories":3165},[],{"categories":3167},[136],{"categories":3169},[183],{"categories":3171},[225],{"categories":3173},[154],{"categories":3175},[90],{"categories":3177},[183],{"categories":3179},[128],{"categories":3181},[],{"categories":3183},[136],{"categories":3185},[90],{"categories":3187},[136],{"categories":3189},[90],{"categories":3191},[183],{"categories":3193},[],{"categories":3195},[136],{"categories":3197},[139],{"categories":3199},[154],{"categories":3201},[136],{"categories":3203},[131],{"categories":3205},[],{"categories":3207},[90],{"categories":3209},[139],{"categories":3211},[90],{"categories":3213},[136],{"categories":3215},[154],{"categories":3217},[128],{"categories":3219},[225],{"categories":3221},[90],{"categories":3223},[90],{"categories":3225},[90],{"categories":3227},[154],{"categories":3229},[131],{"categories":3231},[90],{"categories":3233},[183],{"categories":3235},[154],{"categories":3237},[225],{"categories":3239},[90],{"categories":3241},[],{"categories":3243},[],{"categories":3245},[90],{"categories":3247},[225],{"categories":3249},[186],{"categories":3251},[136],{"categories":3253},[136],{"categories":3255},[154],{"categories":3257},[90],{"categories":3259},[128],{"categories":3261},[183],{"categories":3263},[136],{"categories":3265},[136],{"categories":3267},[90],{"categories":3269},[200],{"categories":3271},[90],{"categories":3273},[136],{"categories":3275},[],{"categories":3277},[90],{"categories":3279},[90],{"categories":3281},[154],{"categories":3283},[128],{"categories":3285},[],{"categories":3287},[90],{"categories":3289},[90],{"categories":3291},[193],{"categories":3293},[183],{"categories":3295},[90,136],{"categories":3297},[200,131],{"categories":3299},[90],{"categories":3301},[],{"categories":3303},[136],{"categories":3305},[],{"categories":3307},[193],{"categories":3309},[90],{"categories":3311},[],{"categories":3313},[90],{"categories":3315},[154],{"categories":3317},[],{"categories":3319},[136],{"categories":3321},[90],{"categories":3323},[],{"categories":3325},[183],{"categories":3327},[136],{"categories":3329},[90],{"categories":3331},[128],{"categories":3333},[136],{"categories":3335},[90],{"categories":3337},[],{"categories":3339},[225],{"categories":3341},[200],{"categories":3343},[131],{"categories":3345},[131],{"categories":3347},[128],{"categories":3349},[128],{"categories":3351},[90],{"categories":3353},[136],{"categories":3355},[90],{"categories":3357},[90],{"categories":3359},[128],{"categories":3361},[90],{"categories":3363},[200],{"categories":3365},[154],{"categories":3367},[90],{"categories":3369},[136],{"categories":3371},[90],{"categories":3373},[],{"categories":3375},[193],{"categories":3377},[],{"categories":3379},[193],{"categories":3381},[136],{"categories":3383},[128],{"categories":3385},[],{"categories":3387},[225],{"categories":3389},[90],{"categories":3391},[],{"categories":3393},[154],{"categories":3395},[136],{"categories":3397},[193],{"categories":3399},[90],{"categories":3401},[136],{"categories":3403},[193],{"categories":3405},[136],{"categories":3407},[154],{"categories":3409},[128],{"categories":3411},[154],{"categories":3413},[193],{"categories":3415},[90],{"categories":3417},[183],{"categories":3419},[90],{"categories":3421},[90],{"categories":3423},[90],{"categories":3425},[90],{"categories":3427},[90],{"categories":3429},[136],{"categories":3431},[90],{"categories":3433},[136],{"categories":3435},[90],{"categories":3437},[128],{"categories":3439},[90],{"categories":3441},[136],{"categories":3443},[183],{"categories":3445},[128],{"categories":3447},[136],{"categories":3449},[183],{"categories":3451},[],{"categories":3453},[90],{"categories":3455},[90],{"categories":3457},[90],{"categories":3459},[193],{"categories":3461},[],{"categories":3463},[136],{"categories":3465},[200],{"categories":3467},[90],{"categories":3469},[154],{"categories":3471},[200],{"categories":3473},[136],{"categories":3475},[131],{"categories":3477},[131],{"categories":3479},[90],{"categories":3481},[90],{"categories":3483},[128],{"categories":3485},[],{"categories":3487},[136],{"categories":3489},[90],{"categories":3491},[],{"categories":3493},[128],{"categories":3495},[90],{"categories":3497},[136],{"categories":3499},[136],{"categories":3501},[],{"categories":3503},[193],{"categories":3505},[193],{"categories":3507},[200],{"categories":3509},[183],{"categories":3511},[],{"categories":3513},[90],{"categories":3515},[136],{"categories":3517},[128],{"categories":3519},[90],{"categories":3521},[193],{"categories":3523},[128],{"categories":3525},[154],{"categories":3527},[154],{"categories":3529},[],{"categories":3531},[154],{"categories":3533},[136],{"categories":3535},[183],{"categories":3537},[186],{"categories":3539},[90],{"categories":3541},[],{"categories":3543},[154],{"categories":3545},[193],{"categories":3547},[131],{"categories":3549},[90],{"categories":3551},[128],{"categories":3553},[225],{"categories":3555},[128],{"categories":3557},[],{"categories":3559},[],{"categories":3561},[154],{"categories":3563},[],{"categories":3565},[136],{"categories":3567},[136],{"categories":3569},[136],{"categories":3571},[],{"categories":3573},[90],{"categories":3575},[],{"categories":3577},[154],{"categories":3579},[128],{"categories":3581},[183],{"categories":3583},[90],{"categories":3585},[154],{"categories":3587},[154],{"categories":3589},[],{"categories":3591},[154],{"categories":3593},[128],{"categories":3595},[136],{"categories":3597},[90],{"categories":3599},[],{"categories":3601},[136],{"categories":3603},[136],{"categories":3605},[128],{"categories":3607},[],{"categories":3609},[],{"categories":3611},[],{"categories":3613},[183],{"categories":3615},[136],{"categories":3617},[90],{"categories":3619},[],{"categories":3621},[],{"categories":3623},[],{"categories":3625},[183],{"categories":3627},[],{"categories":3629},[90],{"categories":3631},[128],{"categories":3633},[],{"categories":3635},[],{"categories":3637},[183],{"categories":3639},[90],{"categories":3641},[154],{"categories":3643},[],{"categories":3645},[200],{"categories":3647},[154],{"categories":3649},[200],{"categories":3651},[186],{"categories":3653},[90],{"categories":3655},[90],{"categories":3657},[],{"categories":3659},[],{"categories":3661},[136],{"categories":3663},[],{"categories":3665},[],{"categories":3667},[136],{"categories":3669},[90],{"categories":3671},[],{"categories":3673},[136],{"categories":3675},[154],{"categories":3677},[90],{"categories":3679},[200],{"categories":3681},[90],{"categories":3683},[186],{"categories":3685},[136],{"categories":3687},[136],{"categories":3689},[],{"categories":3691},[],{"categories":3693},[],{"categories":3695},[154],{"categories":3697},[],{"categories":3699},[],{"categories":3701},[183],{"categories":3703},[128],{"categories":3705},[],{"categories":3707},[131],{"categories":3709},[200],{"categories":3711},[90],{"categories":3713},[193],{"categories":3715},[128],{"categories":3717},[186],{"categories":3719},[131],{"categories":3721},[193],{"categories":3723},[193],{"categories":3725},[],{"categories":3727},[],{"categories":3729},[136],{"categories":3731},[128],{"categories":3733},[183],{"categories":3735},[128],{"categories":3737},[136],{"categories":3739},[225],{"categories":3741},[90],{"categories":3743},[128],{"categories":3745},[136],{"categories":3747},[],{"categories":3749},[90],{"categories":3751},[154],{"categories":3753},[193],{"categories":3755},[],{"categories":3757},[183],{"categories":3759},[154],{"categories":3761},[128],{"categories":3763},[136],{"categories":3765},[90],{"categories":3767},[131],{"categories":3769},[136,225],{"categories":3771},[136],{"categories":3773},[193],{"categories":3775},[90],{"categories":3777},[90],{"categories":3779},[186],{"categories":3781},[200],{"categories":3783},[136],{"categories":3785},[],{"categories":3787},[136],{"categories":3789},[90],{"categories":3791},[131],{"categories":3793},[],{"categories":3795},[],{"categories":3797},[90],{"categories":3799},[186],{"categories":3801},[90],{"categories":3803},[],{"categories":3805},[154],{"categories":3807},[],{"categories":3809},[154],{"categories":3811},[193],{"categories":3813},[128],{"categories":3815},[193],{"categories":3817},[90],{"categories":3819},[136],{"categories":3821},[90],{"categories":3823},[90],{"categories":3825},[200],{"categories":3827},[193],{"categories":3829},[],{"categories":3831},[154],{"categories":3833},[90],{"categories":3835},[],{"categories":3837},[90],{"categories":3839},[90],{"categories":3841},[136],{"categories":3843},[90],{"categories":3845},[136],{"categories":3847},[90],{"categories":3849},[90],{"categories":3851},[90],{"categories":3853},[90],{"categories":3855},[131],{"categories":3857},[],{"categories":3859},[139],{"categories":3861},[154],{"categories":3863},[136],{"categories":3865},[90],{"categories":3867},[],{"categories":3869},[193],{"categories":3871},[90],{"categories":3873},[90],{"categories":3875},[90],{"categories":3877},[136],{"categories":3879},[154],{"categories":3881},[90],{"categories":3883},[90],{"categories":3885},[90],{"categories":3887},[131],{"categories":3889},[90],{"categories":3891},[136],{"categories":3893},[183],{"categories":3895},[],{"categories":3897},[186],{"categories":3899},[90],{"categories":3901},[],{"categories":3903},[154],{"categories":3905},[200],{"categories":3907},[],{"categories":3909},[],{"categories":3911},[154],{"categories":3913},[154],{"categories":3915},[90],{"categories":3917},[200],{"categories":3919},[128],{"categories":3921},[136],{"categories":3923},[90],{"categories":3925},[136],{"categories":3927},[90],{"categories":3929},[131],{"categories":3931},[],{"categories":3933},[186],{"categories":3935},[],{"categories":3937},[154],{"categories":3939},[186],{"categories":3941},[193],{"categories":3943},[136],{"categories":3945},[183],{"categories":3947},[186],{"categories":3949},[186],{"categories":3951},[],{"categories":3953},[154],{"categories":3955},[90],{"categories":3957},[90],{"categories":3959},[193],{"categories":3961},[],{"categories":3963},[154],{"categories":3965},[154],{"categories":3967},[154],{"categories":3969},[],{"categories":3971},[136],{"categories":3973},[90],{"categories":3975},[],{"categories":3977},[128],{"categories":3979},[131],{"categories":3981},[],{"categories":3983},[90],{"categories":3985},[90],{"categories":3987},[],{"categories":3989},[193],{"categories":3991},[],{"categories":3993},[],{"categories":3995},[],{"categories":3997},[],{"categories":3999},[90],{"categories":4001},[154],{"categories":4003},[],{"categories":4005},[],{"categories":4007},[90],{"categories":4009},[90],{"categories":4011},[90],{"categories":4013},[186],{"categories":4015},[90],{"categories":4017},[186],{"categories":4019},[],{"categories":4021},[186],{"categories":4023},[186],{"categories":4025},[225],{"categories":4027},[136],{"categories":4029},[193],{"categories":4031},[],{"categories":4033},[],{"categories":4035},[186],{"categories":4037},[193],{"categories":4039},[193],{"categories":4041},[193],{"categories":4043},[],{"categories":4045},[128],{"categories":4047},[193],{"categories":4049},[193],{"categories":4051},[128],{"categories":4053},[193],{"categories":4055},[131],{"categories":4057},[193],{"categories":4059},[193],{"categories":4061},[193],{"categories":4063},[186],{"categories":4065},[154],{"categories":4067},[154],{"categories":4069},[90],{"categories":4071},[193],{"categories":4073},[186],{"categories":4075},[225],{"categories":4077},[186],{"categories":4079},[186],{"categories":4081},[186],{"categories":4083},[],{"categories":4085},[131],{"categories":4087},[],{"categories":4089},[225],{"categories":4091},[193],{"categories":4093},[193],{"categories":4095},[193],{"categories":4097},[136],{"categories":4099},[154,131],{"categories":4101},[186],{"categories":4103},[],{"categories":4105},[],{"categories":4107},[186],{"categories":4109},[],{"categories":4111},[186],{"categories":4113},[154],{"categories":4115},[136],{"categories":4117},[],{"categories":4119},[193],{"categories":4121},[90],{"categories":4123},[183],{"categories":4125},[],{"categories":4127},[90],{"categories":4129},[],{"categories":4131},[154],{"categories":4133},[128],{"categories":4135},[186],{"categories":4137},[],{"categories":4139},[193],{"categories":4141},[154],[4143,4252,4446,4815],{"id":4144,"title":4145,"ai":4146,"body":4152,"categories":4206,"created_at":91,"date_modified":91,"description":83,"extension":92,"faq":91,"featured":93,"kicker_label":91,"meta":4207,"navigation":106,"path":4239,"published_at":4240,"question":91,"scraped_at":4241,"seo":4242,"sitemap":4243,"source_id":4244,"source_name":4245,"source_type":114,"source_url":4246,"stem":4247,"tags":4248,"thumbnail_url":91,"tldr":4249,"tweet":91,"unknown_tags":4250,"__hash__":4251},"summaries\u002Fsummaries\u002F352a655761b08b28-35b-models-on-rtx-4090-turboquant-kv-compression-u-summary.md","35B Models on RTX 4090: TurboQuant KV Compression Unlocks 32K Context",{"provider":7,"model":4147,"input_tokens":4148,"output_tokens":4149,"processing_time_ms":4150,"cost_usd":4151},"x-ai\u002Fgrok-4.1-fast",6914,2052,13171,0.00190195,{"type":14,"value":4153,"toc":4200},[4154,4158,4161,4165,4168,4172,4178,4184,4190,4193,4197],[17,4155,4157],{"id":4156},"q4_k_m-quantization-delivers-90-95-quality-at-60-original-size","Q4_K_M Quantization Delivers 90-95% Quality at 60% Original Size",[22,4159,4160],{},"Q4_K_M GGUF compresses model weights to ~0.6 GB per billion parameters (7B → 4GB, 32B → 19GB, 70B → 40GB) by storing weights in 4 bits with K-quant block grouping and M-medium mixed precision for sensitive layers. This preserves 90-95% of FP16 accuracy, making it the default for local runs on HuggingFace\u002FOllama. Dense 35B models need 21-22GB VRAM for weights alone on RTX 4090 (24GB total), leaving ~2GB for KV cache—insufficient beyond short contexts. MoE 35B (e.g., Qwen2.5-35B-A3B) activates only 3B params\u002Ftoken, fitting in ~20GB with 1.2GB KV at 64K context due to fewer active heads, reducing TurboQuant's necessity.",[17,4162,4164],{"id":4163},"turboquant-stacks-on-weights-for-long-context-memory-wins","TurboQuant Stacks on Weights for Long-Context Memory Wins",[22,4166,4167],{},"TurboQuant compresses KV cache to 2-4 bits at inference (PolarQuant + QJL, e.g., bits=3) without touching weights, enabling dense models like Mistral Small 3.1 24B or Qwen2.5-32B (64 layers, 8 GQA heads, head_dim=128) to handle 32K context on 24GB VRAM. Formula: 2 × layers × heads × head_dim × seq_len × bytes\u002Felement. Without it, 16K context KV hits ~4GB (total ~24GB borderline); with turbo3, drops to ~1.2GB, freeing space for 32K (~2.4GB). Fused Triton kernels compute attention on compressed KV, speeding up >8K contexts (major at 32K+). Asymmetric K@3bits\u002FV@2bits saves more with zero quality loss empirically.",[17,4169,4171],{"id":4170},"three-paths-to-turboquant-on-24gb-gpus-today","Three Paths to TurboQuant on 24GB GPUs Today",[22,4173,4174,4177],{},[39,4175,4176],{},"PyPI turboquant-kv",": Wrap HF Transformers (load_in_4bit) with TurboQuantModel(bits=3).enable_decoder_fused_attention() for Python scripts; handles 512+ new tokens on long inputs.",[22,4179,4180,4183],{},[39,4181,4182],{},"vLLM fork (0xSero\u002Fturboquant)",": install_turboquant_vllm(bits=3, head_dim=128) before LLM(model, gpu_memory_utilization=0.92); prebuilt codebooks for d=128\u002F256 at 2\u002F3\u002F4 bits; server-friendly.",[22,4185,4186,4189],{},[39,4187,4188],{},"llama.cpp fork (turboquant_plus)",": Build with CUDA, run llama-server -m model-Q4_K_M.gguf --cache-type-k turbo3 --cache-type-v turbo2 -c 32768 -ngl 99. Turbo4 ≈ q8_0 quality, turbo3 best tradeoff, turbo2 extreme. Fits 32K on Qwen2.5-32B (19GB weights + \u003C4GB KV).",[22,4191,4192],{},"Quality holds ≥8B models; speedups context-dependent (\u003C2K: memory only). Experimental—await Google impl (Q2-Q3 2026), llama.cpp #20969, vLLM #38171 merges.",[17,4194,4196],{"id":4195},"optimal-stack-q4_k_m-gguf-turboquant_plus-turbo32","Optimal Stack: Q4_K_M GGUF + turboquant_plus turbo3\u002F2",[22,4198,4199],{},"Download Q4_K_M GGUF, use llama.cpp fork at 16-32K context. Achieves reliable 35B dense inference where defaults crash; 128K impossible (KV still GBs post-compression).",{"title":83,"searchDepth":84,"depth":84,"links":4201},[4202,4203,4204,4205],{"id":4156,"depth":84,"text":4157},{"id":4163,"depth":84,"text":4164},{"id":4170,"depth":84,"text":4171},{"id":4195,"depth":84,"text":4196},[90],{"content_references":4208,"triage":4235},[4209,4214,4217,4221,4223,4226,4229,4232],{"type":4210,"title":4211,"url":4212,"context":4213},"other","GGUF","https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Fhub\u002Fgguf","mentioned",{"type":4210,"title":4215,"url":4216,"context":4213},"AWQ","https:\u002F\u002Fhuggingface.co\u002Fdocs\u002Ftransformers\u002Fquantization\u002Fawq",{"type":4210,"title":4218,"url":4219,"context":4220},"VRAM Requirements for AI Models","https:\u002F\u002Fwillitrunai.com\u002Fblog\u002Fvram-requirements-for-ai-models","cited",{"type":97,"title":4222,"context":100},"turboquant-kv",{"type":97,"title":4224,"url":4225,"context":100},"0xSero\u002Fturboquant","https:\u002F\u002Fgithub.com\u002F0xSero\u002Fturboquant.git",{"type":97,"title":4227,"url":4228,"context":100},"turboquant_plus","https:\u002F\u002Fgithub.com\u002FTheTom\u002Fturboquant_plus.git",{"type":4210,"title":4230,"url":4231,"context":4213},"llama.cpp discussion #20969","https:\u002F\u002Fgithub.com\u002Fggml-org\u002Fllama.cpp\u002Fdiscussions\u002F20969",{"type":4210,"title":4233,"url":4234,"context":4213},"vLLM issue #38171","https:\u002F\u002Fgithub.com\u002Fvllm-project\u002Fvllm\u002Fissues\u002F38171",{"relevance":4236,"novelty":102,"quality":102,"actionability":102,"composite":4237,"reasoning":4238},5,4.35,"Category: AI & LLMs. The article provides in-depth technical insights on running large language models efficiently, addressing the pain point of integrating AI features into products. It offers specific implementation paths for using TurboQuant, which is actionable for developers looking to optimize AI model performance.","\u002Fsummaries\u002F352a655761b08b28-35b-models-on-rtx-4090-turboquant-kv-compression-u-summary","2026-04-15 12:31:01","2026-04-15 15:39:14",{"title":4145,"description":83},{"loc":4239},"352a655761b08b28","Towards AI","https:\u002F\u002Fpub.towardsai.net\u002Frunning-a-35b-model-locally-with-turboquant-whats-actually-possible-right-now-1ac5327430b0?source=rss----98111c9905da---4","summaries\u002F352a655761b08b28-35b-models-on-rtx-4090-turboquant-kv-compression-u-summary",[118,120,119],"Stack Q4_K_M weight quantization with TurboQuant's 3-bit KV cache compression to run dense 35B models at 32K context on 24GB VRAM, fitting weights (20GB) + KV cache (under 4GB) with room to spare—use llama.cpp forks today.",[],"sGRm9blteDamXpWHWdqmPUk1Seq5Nct_nBmywP1bFT0",{"id":4253,"title":4254,"ai":4255,"body":4260,"categories":4425,"created_at":91,"date_modified":91,"description":83,"extension":92,"faq":91,"featured":93,"kicker_label":91,"meta":4426,"navigation":106,"path":4435,"published_at":91,"question":91,"scraped_at":4436,"seo":4437,"sitemap":4438,"source_id":4439,"source_name":4440,"source_type":114,"source_url":4234,"stem":4441,"tags":4442,"thumbnail_url":91,"tldr":4443,"tweet":91,"unknown_tags":4444,"__hash__":4445},"summaries\u002Fsummaries\u002Fd32d038984e0c1db-turboquant-4-7x-kv-cache-compression-in-vllm-summary.md","TurboQuant: 4-7x KV Cache Compression in vLLM",{"provider":7,"model":4147,"input_tokens":4256,"output_tokens":4257,"processing_time_ms":4258,"cost_usd":4259},10176,1474,8441,0.0027497,{"type":14,"value":4261,"toc":4420},[4262,4266,4269,4272,4276,4279,4358,4361,4365,4368,4417],[17,4263,4265],{"id":4264},"turboquant-delivers-superior-kv-cache-compression","TurboQuant Delivers Superior KV Cache Compression",[22,4267,4268],{},"TurboQuant uses online vector quantization with QR rotation, Lloyd-Max codebooks, and bit-packing for 2-4 bit (including 2.5\u002F3.5 fractional) KV caches, achieving provably near-optimal distortion within 2.7x of information-theoretic limits. Unlike scalar methods like FP8 (e4m3\u002Fe5m2) or INT4, it preserves inner products unbiased—key for attention—while enabling 4-5x memory savings. Paper benchmarks show perfect Needle-in-a-Haystack recall at 4x compression and competitive LongBench scores at 2.5-3.5 bits\u002Fdim. It requires no preprocessing, runs online, and suits accelerators.",[22,4270,4271],{},"vLLM alternatives (FP8, compressed-tensors) optimize MSE element-wise but lack vector codebooks, inner-product focus, theoretical guarantees, or sub-4-bit flexibility.",[17,4273,4275],{"id":4274},"proven-zero-loss-performance-and-throughput-gains","Proven Zero-Loss Performance and Throughput Gains",[22,4277,4278],{},"PoC on Qwen2.5-7B (H200, 4K-16K context) yields:",[4280,4281,4282,4301],"table",{},[4283,4284,4285],"thead",{},[4286,4287,4288,4292,4295,4298],"tr",{},[4289,4290,4291],"th",{},"Config",[4289,4293,4294],{},"Exact Match",[4289,4296,4297],{},"Avg Cache GB",[4289,4299,4300],{},"vs Full",[4302,4303,4304,4319,4332,4345],"tbody",{},[4286,4305,4306,4310,4313,4316],{},[4307,4308,4309],"td",{},"Full",[4307,4311,4312],{},"6\u002F6",[4307,4314,4315],{},"0.510",[4307,4317,4318],{},"1.0x",[4286,4320,4321,4324,4326,4329],{},[4307,4322,4323],{},"TQ 2-bit",[4307,4325,4312],{},[4307,4327,4328],{},"0.068",[4307,4330,4331],{},"7.5x",[4286,4333,4334,4337,4339,4342],{},[4307,4335,4336],{},"TQ 3.5-bit",[4307,4338,4312],{},[4307,4340,4341],{},"0.112",[4307,4343,4344],{},"4.5x",[4286,4346,4347,4350,4352,4355],{},[4307,4348,4349],{},"TQ 4-bit",[4307,4351,4312],{},[4307,4353,4354],{},"0.132",[4307,4356,4357],{},"3.9x",[22,4359,4360],{},"Upstream PR #38280 (Qwen2.5-1.5B, H200) confirms 12\u002F12 exact matches across bit-widths, TTFT\u002FITL latency matching baseline (9.3ms\u002F8.4ms), and 21% throughput boost at batch=16. Phase 2 adds bit-packed uint8 storage (ceil(head_size*bits\u002F8)+2 bytes\u002Fslot) for full ratios.",[17,4362,4364],{"id":4363},"straightforward-vllm-integration-path","Straightforward vLLM Integration Path",[22,4366,4367],{},"Aligns with vLLM's framework:",[33,4369,4370,4386,4393,4400,4407,4410],{},[36,4371,4372,4373,4377,4378,4381,4382,4385],{},"Extend ",[4374,4375,4376],"code",{},"CacheDType"," in ",[4374,4379,4380],{},"cache.py","\u002F",[4374,4383,4384],{},"torch_utils.py"," for integer indices.",[36,4387,4388,4389,4392],{},"Add ",[4374,4390,4391],{},"@register_quantization_config(\"turboquant\") TurboQuantConfig"," targeting Attention layers.",[36,4394,4395,4396,4399],{},"Implement ",[4374,4397,4398],{},"TurboQuantKVCacheMethod"," (extends BaseKVCacheMethod) for codebook params, MSE\u002FIP variants, per-head support.",[36,4401,4402,4403,4406],{},"Update ",[4374,4404,4405],{},"is_quantized_kv_cache()"," detection.",[36,4408,4409],{},"CUDA\u002FTriton encode\u002Fdecode kernels (43\u002F43 tests pass).",[36,4411,4412,4413,4416],{},"Adjust ",[4374,4414,4415],{},"KVCacheSpec"," for codebook overhead\u002Fvariable ratios.",[22,4418,4419],{},"PoC covers steps 1-5; PR #38280 integrates fully with Triton attention. Related: PolarQuant, ollama\u002Follama#15051, llama.cpp#20977, vllm-omni#2214.",{"title":83,"searchDepth":84,"depth":84,"links":4421},[4422,4423,4424],{"id":4264,"depth":84,"text":4265},{"id":4274,"depth":84,"text":4275},{"id":4363,"depth":84,"text":4364},[90],{"content_references":4427,"triage":4432},[4428],{"type":4429,"title":4430,"url":4431,"context":4220},"paper","TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate","https:\u002F\u002Farxiv.org\u002Fpdf\u002F2504.19874",{"relevance":102,"novelty":103,"quality":102,"actionability":102,"composite":4433,"reasoning":4434},3.8,"Category: AI & LLMs. The article discusses TurboQuant's vector quantization for KV cache compression, which is relevant for AI engineers looking to optimize LLM performance. It provides specific integration steps for vLLM, making it actionable for developers, though the content is quite technical and may not be accessible to all audiences.","\u002Fsummaries\u002Fd32d038984e0c1db-turboquant-4-7x-kv-cache-compression-in-vllm-summary","2026-04-16 03:08:39",{"title":4254,"description":83},{"loc":4435},"d32d038984e0c1db","__oneoff__","summaries\u002Fd32d038984e0c1db-turboquant-4-7x-kv-cache-compression-in-vllm-summary",[118,120,119],"TurboQuant vector quantization compresses vLLM KV caches 3.9-7.5x at 2-4 bits\u002Fdim with perfect Needle-in-a-Haystack recall, zero latency overhead, and 21% throughput gains.",[],"XlnqSgcG5nhgzzRW6Sdk3XPgtmFuSj6gJC1hWx2uIwk",{"id":4447,"title":4448,"ai":4449,"body":4454,"categories":4801,"created_at":91,"date_modified":91,"description":83,"extension":92,"faq":91,"featured":93,"kicker_label":91,"meta":4802,"navigation":106,"path":4803,"published_at":4804,"question":91,"scraped_at":91,"seo":4805,"sitemap":4806,"source_id":4807,"source_name":4808,"source_type":114,"source_url":4809,"stem":4810,"tags":4811,"thumbnail_url":91,"tldr":4812,"tweet":91,"unknown_tags":4813,"__hash__":4814},"summaries\u002Fsummaries\u002Fllm-as-judge-evaluates-rag-keyword-beats-vector-summary.md","LLM-as-Judge Evaluates RAG: Keyword Beats Vector",{"provider":7,"model":4147,"input_tokens":4450,"output_tokens":4451,"processing_time_ms":4452,"cost_usd":4453},5849,1975,17506,0.0021348,{"type":14,"value":4455,"toc":4796},[4456,4460,4467,4473,4493,4498,4516,4531,4535,4546,4567,4570,4641,4644,4671,4674,4694,4697,4732,4736,4739,4785,4792],[17,4457,4459],{"id":4458},"rag-needs-automated-internal-evaluation-for-optimization","RAG Needs Automated Internal Evaluation for Optimization",[22,4461,4462,4463,4466],{},"RAG systems require quantitative evaluation to compare optimizations like retrieval strategies, avoiding manual checks that are slow and subjective—integrate into CI\u002FCD pipelines like unit tests. Focus on ",[39,4464,4465],{},"internal evaluation"," of retrieval and generation modules:",[22,4468,4469,4472],{},[39,4470,4471],{},"Retrieval metrics",":",[33,4474,4475,4481,4487],{},[36,4476,4477,4480],{},[39,4478,4479],{},"Relevance",": Retrieved chunks match query?",[36,4482,4483,4486],{},[39,4484,4485],{},"Coverage",": All relevant database chunks fetched?",[36,4488,4489,4492],{},[39,4490,4491],{},"Correctness",": High signal-to-noise ratio, relevant chunks ranked top?",[22,4494,4495,4472],{},[39,4496,4497],{},"Generation metrics",[33,4499,4500,4505,4511],{},[36,4501,4502,4504],{},[39,4503,4479],{},": Answer aligns with query, no off-topic drift?",[36,4506,4507,4510],{},[39,4508,4509],{},"Factuality",": Answer grounded in retrieved sources, no hallucinations?",[36,4512,4513,4515],{},[39,4514,4491],{},": Answer factually accurate?",[22,4517,4518,4519,4522,4523,4526,4527,4530],{},"Prefer ",[39,4520,4521],{},"LLM-as-a-judge"," over traditional NLP metrics (ROUGE, BLEU) for nuanced semantic judgment. Ground evaluators in production setups like Azure AI Search indexes (e.g., ",[4374,4524,4525],{},"rag-evalution-chris"," with 50 chunks from employee handbook PDFs, vectorized in ",[4374,4528,4529],{},"text_vector"," field).",[17,4532,4534],{"id":4533},"azure-sdk-evaluators-automate-llm-as-judge-scoring","Azure SDK Evaluators Automate LLM-as-Judge Scoring",[22,4536,4537,4538,4541,4542,4545],{},"Leverage ",[4374,4539,4540],{},"azure.ai.evaluation"," package with GPT-4 (",[4374,4543,4544],{},"gpt-4.1"," deployment) for zero-shot scoring (1.0-5.0 scale). Key evaluators:",[33,4547,4548,4558],{},[36,4549,4550,4553,4554,4557],{},[39,4551,4552],{},"GroundednessEvaluator",": Measures answer's fidelity to sources—scores drop if facts can't be verified in context, even if externally true. Input: ",[4374,4555,4556],{},"response=answer, context=sources",".",[36,4559,4560,4563,4564,4557],{},[39,4561,4562],{},"RelevanceEvaluator",": Checks query-response alignment and contextual fit. Input: ",[4374,4565,4566],{},"query=user_question, response=answer, context=sources",[22,4568,4569],{},"Setup clients for Azure AI Search and OpenAI:",[4571,4572,4575],"pre",{"className":4573,"code":4574,"language":119,"meta":83,"style":83},"language-python shiki shiki-themes github-light github-dark","import os\nfrom azure.search.documents import SearchClient\nfrom azure.search.documents.models import VectorizedQuery\nfrom openai import AzureOpenAI\n# Load env vars: AZURE_SEARCH_*, AZURE_OPENAI_*\nopenai_client = AzureOpenAI(api_key=AZURE_OPENAI_API_KEY, azure_endpoint=AZURE_OPENAI_ENDPOINT, api_version=\"2024-10-21\")\nsearch_client = SearchClient(endpoint=AZURE_SEARCH_ENDPOINT, index_name=AZURE_SEARCH_INDEX_NAME, credential=AzureKeyCredential(AZURE_SEARCH_ADMIN_KEY))\n\ndef get_embedding_vector(query: str) -> list[float]:\n    response = openai_client.embeddings.create(model=AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME, input=[query])\n    return response.data[0].embedding\n",[4374,4576,4577,4585,4590,4595,4600,4605,4611,4617,4623,4629,4635],{"__ignoreMap":83},[4578,4579,4582],"span",{"class":4580,"line":4581},"line",1,[4578,4583,4584],{},"import os\n",[4578,4586,4587],{"class":4580,"line":84},[4578,4588,4589],{},"from azure.search.documents import SearchClient\n",[4578,4591,4592],{"class":4580,"line":103},[4578,4593,4594],{},"from azure.search.documents.models import VectorizedQuery\n",[4578,4596,4597],{"class":4580,"line":102},[4578,4598,4599],{},"from openai import AzureOpenAI\n",[4578,4601,4602],{"class":4580,"line":4236},[4578,4603,4604],{},"# Load env vars: AZURE_SEARCH_*, AZURE_OPENAI_*\n",[4578,4606,4608],{"class":4580,"line":4607},6,[4578,4609,4610],{},"openai_client = AzureOpenAI(api_key=AZURE_OPENAI_API_KEY, azure_endpoint=AZURE_OPENAI_ENDPOINT, api_version=\"2024-10-21\")\n",[4578,4612,4614],{"class":4580,"line":4613},7,[4578,4615,4616],{},"search_client = SearchClient(endpoint=AZURE_SEARCH_ENDPOINT, index_name=AZURE_SEARCH_INDEX_NAME, credential=AzureKeyCredential(AZURE_SEARCH_ADMIN_KEY))\n",[4578,4618,4620],{"class":4580,"line":4619},8,[4578,4621,4622],{"emptyLinePlaceholder":106},"\n",[4578,4624,4626],{"class":4580,"line":4625},9,[4578,4627,4628],{},"def get_embedding_vector(query: str) -> list[float]:\n",[4578,4630,4632],{"class":4580,"line":4631},10,[4578,4633,4634],{},"    response = openai_client.embeddings.create(model=AZURE_OPENAI_EMBEDDING_DEPLOYMENT_NAME, input=[query])\n",[4578,4636,4638],{"class":4580,"line":4637},11,[4578,4639,4640],{},"    return response.data[0].embedding\n",[22,4642,4643],{},"Retrieval (top=5):",[33,4645,4646,4655,4663],{},[36,4647,4648,4651,4652],{},[39,4649,4650],{},"Keyword",": ",[4374,4653,4654],{},"search_client.search(search_text=user_question)",[36,4656,4657,4651,4660],{},[39,4658,4659],{},"Vector",[4374,4661,4662],{},"search_client.search(None, vector_queries=[VectorizedQuery(vector=get_embedding_vector(user_question), k_nearest_neighbors=50, fields=\"text_vector\")])",[36,4664,4665,4651,4668],{},[39,4666,4667],{},"Hybrid (semantic)",[4374,4669,4670],{},"search_client.search(user_question, vector_queries=[...], query_type=\"semantic\", semantic_configuration_name=\"rag-evaluation-chris-semantic-configuration\")",[22,4672,4673],{},"Generation prompt enforces grounding:",[4571,4675,4677],{"className":4573,"code":4676,"language":119,"meta":83,"style":83},"SYSTEM_MESSAGE = \"\"\"Answer ONLY with facts from sources. Use [source] citations.\"\"\"\nresponse = openai_client.chat.completions.create(model=AZURE_OPENAI_LLM_DEPLOYMENT_NAME, messages=[{\"role\": \"system\", \"content\": SYSTEM_MESSAGE}, {\"role\": \"user\", \"content\": user_question + \"\\nSources: \" + sources}])\nanswer = response.choices[0].message.content\n",[4374,4678,4679,4684,4689],{"__ignoreMap":83},[4578,4680,4681],{"class":4580,"line":4581},[4578,4682,4683],{},"SYSTEM_MESSAGE = \"\"\"Answer ONLY with facts from sources. Use [source] citations.\"\"\"\n",[4578,4685,4686],{"class":4580,"line":84},[4578,4687,4688],{},"response = openai_client.chat.completions.create(model=AZURE_OPENAI_LLM_DEPLOYMENT_NAME, messages=[{\"role\": \"system\", \"content\": SYSTEM_MESSAGE}, {\"role\": \"user\", \"content\": user_question + \"\\nSources: \" + sources}])\n",[4578,4690,4691],{"class":4580,"line":103},[4578,4692,4693],{},"answer = response.choices[0].message.content\n",[22,4695,4696],{},"Evaluate:",[4571,4698,4700],{"className":4573,"code":4699,"language":119,"meta":83,"style":83},"from azure.ai.evaluation import AzureOpenAIModelConfiguration, GroundednessEvaluator, RelevanceEvaluator\nmodel_config = {\"azure_endpoint\": AZURE_OPENAI_ENDPOINT, \"azure_deployment\": AZURE_OPENAI_LLM_DEPLOYMENT_NAME, \"api_key\": AZURE_OPENAI_API_KEY}\nrelevance_eval = RelevanceEvaluator(model_config)\ngroundedness_eval = GroundednessEvaluator(model_config)\nrelevance_score = relevance_eval(query=user_question, response=answer, context=sources)\ngroundedness_score = groundedness_eval(response=answer, context=sources)\n",[4374,4701,4702,4707,4712,4717,4722,4727],{"__ignoreMap":83},[4578,4703,4704],{"class":4580,"line":4581},[4578,4705,4706],{},"from azure.ai.evaluation import AzureOpenAIModelConfiguration, GroundednessEvaluator, RelevanceEvaluator\n",[4578,4708,4709],{"class":4580,"line":84},[4578,4710,4711],{},"model_config = {\"azure_endpoint\": AZURE_OPENAI_ENDPOINT, \"azure_deployment\": AZURE_OPENAI_LLM_DEPLOYMENT_NAME, \"api_key\": AZURE_OPENAI_API_KEY}\n",[4578,4713,4714],{"class":4580,"line":103},[4578,4715,4716],{},"relevance_eval = RelevanceEvaluator(model_config)\n",[4578,4718,4719],{"class":4580,"line":102},[4578,4720,4721],{},"groundedness_eval = GroundednessEvaluator(model_config)\n",[4578,4723,4724],{"class":4580,"line":4236},[4578,4725,4726],{},"relevance_score = relevance_eval(query=user_question, response=answer, context=sources)\n",[4578,4728,4729],{"class":4580,"line":4607},[4578,4730,4731],{},"groundedness_score = groundedness_eval(response=answer, context=sources)\n",[17,4733,4735],{"id":4734},"keyword-search-wins-for-simple-queries-enables-agentic-rag","Keyword Search Wins for Simple Queries, Enables Agentic RAG",[22,4737,4738],{},"On query \"What does a product manager do?\" (50-chunk index):",[4280,4740,4741,4753],{},[4283,4742,4743],{},[4286,4744,4745,4748,4751],{},[4289,4746,4747],{},"Method",[4289,4749,4750],{},"Groundedness",[4289,4752,4479],{},[4302,4754,4755,4765,4775],{},[4286,4756,4757,4759,4762],{},[4307,4758,4650],{},[4307,4760,4761],{},"4.5",[4307,4763,4764],{},"5.0",[4286,4766,4767,4770,4773],{},[4307,4768,4769],{},"Hybrid",[4307,4771,4772],{},"4.0",[4307,4774,4761],{},[4286,4776,4777,4779,4782],{},[4307,4778,4659],{},[4307,4780,4781],{},"3.0",[4307,4783,4784],{},"3.5",[22,4786,4787,4788,4791],{},"Keyword search topped scores unexpectedly for this task, proving automated eval reveals trade-offs (e.g., vector struggles with exact phrasing). This closes the loop for ",[39,4789,4790],{},"Agentic RAG",": reliable metrics select best retrieval for self-improving agents.",[4793,4794,4795],"style",{},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"title":83,"searchDepth":84,"depth":84,"links":4797},[4798,4799,4800],{"id":4458,"depth":84,"text":4459},{"id":4533,"depth":84,"text":4534},{"id":4734,"depth":84,"text":4735},[],{},"\u002Fsummaries\u002Fllm-as-judge-evaluates-rag-keyword-beats-vector-summary","2026-04-08 21:21:17",{"title":4448,"description":83},{"loc":4803},"20b8d035b68db639","Level Up Coding","https:\u002F\u002Funknown","summaries\u002Fllm-as-judge-evaluates-rag-keyword-beats-vector-summary",[118,119,120],"Use Azure SDK's GroundednessEvaluator (1-5 scale: answer fidelity to sources) and RelevanceEvaluator (query-response alignment) to automate RAG scoring; keyword search outperformed vector\u002Fhybrid on 'product manager duties' query.",[],"WbpHbWCJqCpSbai5hGyHaCpcDDTRGieYViPox5fj2mc",{"id":4816,"title":4817,"ai":4818,"body":4823,"categories":4904,"created_at":91,"date_modified":91,"description":83,"extension":92,"faq":91,"featured":93,"kicker_label":91,"meta":4905,"navigation":106,"path":4929,"published_at":91,"question":91,"scraped_at":4930,"seo":4931,"sitemap":4932,"source_id":4933,"source_name":4440,"source_type":114,"source_url":4934,"stem":4935,"tags":4936,"thumbnail_url":91,"tldr":4937,"tweet":91,"unknown_tags":4938,"__hash__":4939},"summaries\u002Fsummaries\u002F8fee41411642a9b7-harmony-render-gpt-oss-response-format-in-rust-pyt-summary.md","Harmony: Render gpt-oss Response Format in Rust\u002FPython",{"provider":7,"model":4147,"input_tokens":4819,"output_tokens":4820,"processing_time_ms":4821,"cost_usd":4822},5558,1823,8018,0.00151625,{"type":14,"value":4824,"toc":4899},[4825,4829,4832,4835,4843,4846,4850,4861,4864,4868],[17,4826,4828],{"id":4827},"harmony-format-enables-structured-gpt-oss-outputs","Harmony Format Enables Structured gpt-oss Outputs",[22,4830,4831],{},"gpt-oss models demand the harmony response format for correct operation, as they were trained specifically on it. This format structures conversations, reasoning traces, and function calls using special tokens like \u003C|start|>role\u003C|message|> and \u003C|end|>. It supports multiple channels (e.g., analysis, commentary, final) for separating chain-of-thought, tool preambles, and responses, plus tool namespaces and structured outputs with instruction hierarchies. Without harmony, gpt-oss fails; providers like HuggingFace, Ollama, or vLLM handle it automatically, but custom inference requires manual prompting.",[22,4833,4834],{},"Example system prompt specifies channels and tools:",[4571,4836,4841],{"className":4837,"code":4839,"language":4840},[4838],"language-text","\u003C|start|>system\u003C|message|>You are ChatGPT... Reasoning: high # Valid channels: analysis, commentary, final... Calls to 'functions' must go to commentary.\u003C|end|>\n\u003C|start|>developer\u003C|message|># Instructions Always respond in riddles # Tools ## functions namespace functions { type get_location = () => any; type get_current_weather = (_: {location: string...}) => any; }\u003C|end|>\n\u003C|start|>user\u003C|message|>What is the weather like in SF?\u003C|end|>\n\u003C|start|>assistant\n","text",[4374,4842,4839],{"__ignoreMap":83},[22,4844,4845],{},"This mimics OpenAI's Responses API, easing transition for familiar users. See full guide at cookbook.openai.com\u002Farticles\u002Fopenai-harmony.",[17,4847,4849],{"id":4848},"python-and-rust-libraries-for-encodingparsing","Python and Rust Libraries for Encoding\u002FParsing",[22,4851,4852,4853,4856,4857,4860],{},"Install Python via ",[4374,4854,4855],{},"pip install openai-harmony"," for high-level dataclasses mirroring chat structures (Role, Message). Rust core handles rendering\u002Fparsing via ",[4374,4858,4859],{},"cargo add harmony",", with full docs at docs\u002Fpython.md and docs\u002Frust.md.",[22,4862,4863],{},"Architecture: Rust crate (src\u002F) with chat.rs for data structures, encoding.rs for logic, tiktoken tokenizer, and registry.rs for encodings. Python wrapper (python\u002Fopenai_harmony\u002F) uses pyo3 FFI bindings, producing openai_harmony.*.so. Repo includes tests\u002F, test-data\u002F, demo\u002Fharmony-demo, and AGENTS.md.",[17,4865,4867],{"id":4866},"local-development-and-testing","Local Development and Testing",[22,4869,4870,4871,4874,4875,4878,4879,4882,4883,4886,4887,4890,4891,4894,4895,4898],{},"Use ",[4374,4872,4873],{},"maturin develop"," to build Rust extension into virtualenv, then ",[4374,4876,4877],{},"pip install -e ."," for Python wrapper. Run Rust tests with ",[4374,4880,4881],{},"cargo test",", Python with ",[4374,4884,4885],{},"pytest tests",", or both via ",[4374,4888,4889],{},".\u002Frun_checks.sh",". Optional: ",[4374,4892,4893],{},"cargo fmt",", ",[4374,4896,4897],{},"ruff check .",". Ensures Rust\u002FPython parity for performance-critical rendering.",{"title":83,"searchDepth":84,"depth":84,"links":4900},[4901,4902,4903],{"id":4827,"depth":84,"text":4828},{"id":4848,"depth":84,"text":4849},{"id":4866,"depth":84,"text":4867},[],{"content_references":4906,"triage":4926},[4907,4910,4912,4915,4918,4921,4924],{"type":97,"title":4908,"url":4909,"context":4213},"gpt-oss","https:\u002F\u002Fopenai.com\u002Fopen-models",{"type":97,"title":4908,"url":4911,"context":100},"https:\u002F\u002Fgpt-oss.com",{"type":4210,"title":4913,"url":4914,"context":4213},"gpt-oss Model Card","https:\u002F\u002Fopenai.com\u002Findex\u002Fgpt-oss-model-card\u002F",{"type":4210,"title":4916,"url":4917,"context":4220},"OpenAI Harmony Guide","https:\u002F\u002Fcookbook.openai.com\u002Farticles\u002Fopenai-harmony",{"type":4210,"title":4919,"url":4920,"context":4213},"gpt-oss Cookbook","https:\u002F\u002Fcookbook.openai.com\u002Ftopic\u002Fgpt-oss",{"type":97,"title":4922,"url":4923,"context":4213},"pyo3","https:\u002F\u002Fpyo3.rs\u002F",{"type":97,"title":4925,"context":4213},"maturin",{"relevance":4236,"novelty":103,"quality":102,"actionability":102,"composite":4927,"reasoning":4928},4.15,"Category: AI & LLMs. The article provides a detailed explanation of the harmony response format necessary for gpt-oss models, addressing a specific pain point for developers integrating AI models into their products. It includes practical installation instructions and examples, making it actionable for developers.","\u002Fsummaries\u002F8fee41411642a9b7-harmony-render-gpt-oss-response-format-in-rust-pyt-summary","2026-04-16 03:07:33",{"title":4817,"description":83},{"loc":4929},"8fee41411642a9b7","https:\u002F\u002Fgithub.com\u002Fopenai\u002Fharmony","summaries\u002F8fee41411642a9b7-harmony-render-gpt-oss-response-format-in-rust-pyt-summary",[118,120,119],"OpenAI's harmony library encodes\u002Fdecodes the harmony response format required for gpt-oss open-weight models in custom inference setups, mimicking the OpenAI API with multi-channel support for reasoning and tools.",[],"C5U8PWt3Bm5l1z20k8w-gzxDIGVbaFty5WmLaRh64O0"]