45c3aad453
- Add Claude Opus 4.7, Kimi K2.6, GLM-5.1 to existing GLM-5, Qwen3-6, MiniMax-M2.7 - Add 5 new challenges: flash attention fwd/bwd, beam search, DFlash, ternary training - Rewrite README with TL;DR rankings, grade matrix, and DeepSeek V4 Pro attribution - Add analysis/ folder with cross-model comparisons and per-challenge deep dives - Add deploy_challenges.sh script - Expand .gitignore to exclude Python envs, ML weights, and build artifacts
520 lines
13 KiB
JSON
520 lines
13 KiB
JSON
{
|
|
"hyperparameters": {
|
|
"group_size": 128,
|
|
"seq_length": 128,
|
|
"batch_size": 2,
|
|
"num_steps": 500,
|
|
"learning_rate": 5e-05
|
|
},
|
|
"training": {
|
|
"initial_loss": 19.415189743041992,
|
|
"final_loss": 2.9524176120758057,
|
|
"loss_curve": [
|
|
19.415189743041992,
|
|
17.6638126373291,
|
|
17.5764102935791,
|
|
15.895140647888184,
|
|
14.59153938293457,
|
|
16.177194595336914,
|
|
12.788802146911621,
|
|
11.541013717651367,
|
|
12.024927139282227,
|
|
12.84988021850586,
|
|
11.819856643676758,
|
|
11.076229095458984,
|
|
10.465115547180176,
|
|
10.294048309326172,
|
|
9.291016578674316,
|
|
5.1339216232299805,
|
|
5.194877624511719,
|
|
10.74880313873291,
|
|
13.278043746948242,
|
|
10.388489723205566,
|
|
10.168744087219238,
|
|
10.00532341003418,
|
|
10.9448881149292,
|
|
13.310420989990234,
|
|
11.944724082946777,
|
|
8.91694450378418,
|
|
5.131122589111328,
|
|
10.197563171386719,
|
|
10.01301097869873,
|
|
8.23267936706543,
|
|
10.079784393310547,
|
|
9.714762687683105,
|
|
9.727270126342773,
|
|
7.821086406707764,
|
|
7.828431606292725,
|
|
8.294801712036133,
|
|
8.915307998657227,
|
|
6.78751802444458,
|
|
8.784393310546875,
|
|
7.263306617736816,
|
|
6.968741416931152,
|
|
7.635254859924316,
|
|
8.597051620483398,
|
|
3.962278366088867,
|
|
2.649355888366699,
|
|
1.3925496339797974,
|
|
2.150885581970215,
|
|
2.4210667610168457,
|
|
2.107257127761841,
|
|
2.2125606536865234,
|
|
2.3535611629486084,
|
|
2.306110382080078,
|
|
2.496791362762451,
|
|
1.910773515701294,
|
|
2.7408840656280518,
|
|
2.522926092147827,
|
|
2.7258849143981934,
|
|
2.2239017486572266,
|
|
7.865502834320068,
|
|
11.369744300842285,
|
|
7.2785325050354,
|
|
10.744400024414062,
|
|
11.363978385925293,
|
|
11.373944282531738,
|
|
11.010241508483887,
|
|
9.836434364318848,
|
|
9.70583438873291,
|
|
9.144129753112793,
|
|
9.465516090393066,
|
|
8.564160346984863,
|
|
9.16439151763916,
|
|
9.109588623046875,
|
|
8.336548805236816,
|
|
7.943811893463135,
|
|
7.852457523345947,
|
|
6.610472679138184,
|
|
4.080750465393066,
|
|
4.603875637054443,
|
|
7.214062690734863,
|
|
14.470048904418945,
|
|
14.226180076599121,
|
|
13.72647762298584,
|
|
10.721973419189453,
|
|
10.570021629333496,
|
|
10.210411071777344,
|
|
10.290371894836426,
|
|
7.791189193725586,
|
|
7.8287835121154785,
|
|
7.902010440826416,
|
|
8.494746208190918,
|
|
8.886126518249512,
|
|
8.344682693481445,
|
|
9.480537414550781,
|
|
8.99856948852539,
|
|
8.164642333984375,
|
|
8.365951538085938,
|
|
9.024402618408203,
|
|
8.6676607131958,
|
|
10.06509017944336,
|
|
9.371912956237793,
|
|
9.17201042175293,
|
|
9.499948501586914,
|
|
8.475625991821289,
|
|
9.137506484985352,
|
|
8.084639549255371,
|
|
8.213334083557129,
|
|
7.3555707931518555,
|
|
7.324641227722168,
|
|
7.4844536781311035,
|
|
8.139140129089355,
|
|
7.955804824829102,
|
|
8.107175827026367,
|
|
6.985445022583008,
|
|
6.115233421325684,
|
|
6.798851013183594,
|
|
2.756054639816284,
|
|
4.928526401519775,
|
|
9.184700012207031,
|
|
9.650903701782227,
|
|
7.893393039703369,
|
|
7.769137382507324,
|
|
7.712228775024414,
|
|
8.659494400024414,
|
|
9.301843643188477,
|
|
9.03166675567627,
|
|
7.267263889312744,
|
|
8.050270080566406,
|
|
8.89819049835205,
|
|
7.454459190368652,
|
|
7.789579391479492,
|
|
8.938220977783203,
|
|
8.343205451965332,
|
|
7.659829616546631,
|
|
7.563717842102051,
|
|
7.64760160446167,
|
|
6.753893852233887,
|
|
7.2767486572265625,
|
|
7.687180042266846,
|
|
8.177096366882324,
|
|
5.205698013305664,
|
|
8.55665397644043,
|
|
8.401761054992676,
|
|
8.025993347167969,
|
|
8.522932052612305,
|
|
7.386404514312744,
|
|
6.299332141876221,
|
|
7.9422607421875,
|
|
6.485499382019043,
|
|
7.92954158782959,
|
|
5.921761512756348,
|
|
7.883401870727539,
|
|
7.638513088226318,
|
|
7.558638095855713,
|
|
7.362685203552246,
|
|
8.297099113464355,
|
|
8.487621307373047,
|
|
8.52571964263916,
|
|
8.659907341003418,
|
|
8.015156745910645,
|
|
9.298934936523438,
|
|
8.222744941711426,
|
|
6.188640594482422,
|
|
8.977818489074707,
|
|
8.637101173400879,
|
|
8.659961700439453,
|
|
7.4918599128723145,
|
|
8.798979759216309,
|
|
7.740288257598877,
|
|
8.463373184204102,
|
|
8.464582443237305,
|
|
7.778406620025635,
|
|
9.147701263427734,
|
|
7.360451698303223,
|
|
7.708859443664551,
|
|
6.682768821716309,
|
|
7.512155055999756,
|
|
8.024608612060547,
|
|
8.361748695373535,
|
|
5.732519149780273,
|
|
6.673101425170898,
|
|
7.6330132484436035,
|
|
8.132368087768555,
|
|
7.8759942054748535,
|
|
8.514373779296875,
|
|
8.397266387939453,
|
|
7.0031304359436035,
|
|
7.621158123016357,
|
|
7.67484188079834,
|
|
7.817298889160156,
|
|
7.450564861297607,
|
|
6.986921310424805,
|
|
9.063298225402832,
|
|
7.272268772125244,
|
|
8.928145408630371,
|
|
6.965574264526367,
|
|
9.52602767944336,
|
|
7.277902126312256,
|
|
6.177265167236328,
|
|
8.317046165466309,
|
|
8.4580078125,
|
|
8.824596405029297,
|
|
7.85051965713501,
|
|
5.829211711883545,
|
|
8.68645191192627,
|
|
8.018779754638672,
|
|
7.682953834533691,
|
|
8.003823280334473,
|
|
6.92888879776001,
|
|
6.7287917137146,
|
|
7.22535514831543,
|
|
6.919946670532227,
|
|
7.498782634735107,
|
|
7.409185409545898,
|
|
8.3101167678833,
|
|
6.284835338592529,
|
|
3.541412115097046,
|
|
3.9863815307617188,
|
|
6.179129123687744,
|
|
6.740180492401123,
|
|
7.888493537902832,
|
|
4.698310852050781,
|
|
5.089892864227295,
|
|
8.01733112335205,
|
|
4.149894714355469,
|
|
3.0928893089294434,
|
|
9.866519927978516,
|
|
8.222246170043945,
|
|
5.943643569946289,
|
|
8.004118919372559,
|
|
5.507823944091797,
|
|
8.96957015991211,
|
|
6.324719429016113,
|
|
8.650246620178223,
|
|
8.170387268066406,
|
|
8.473105430603027,
|
|
8.394067764282227,
|
|
5.1943159103393555,
|
|
3.4560070037841797,
|
|
2.6845388412475586,
|
|
2.9381015300750732,
|
|
8.991165161132812,
|
|
9.567828178405762,
|
|
9.947354316711426,
|
|
6.080748081207275,
|
|
5.691708564758301,
|
|
7.181239604949951,
|
|
8.073373794555664,
|
|
8.77186107635498,
|
|
8.518348693847656,
|
|
7.958341598510742,
|
|
8.752128601074219,
|
|
7.485937595367432,
|
|
8.58120346069336,
|
|
8.627962112426758,
|
|
6.968264102935791,
|
|
7.434549331665039,
|
|
7.358287334442139,
|
|
7.684825897216797,
|
|
7.424722194671631,
|
|
6.908591270446777,
|
|
6.278493404388428,
|
|
8.345937728881836,
|
|
7.803347587585449,
|
|
8.391436576843262,
|
|
8.13833999633789,
|
|
8.466653823852539,
|
|
8.621729850769043,
|
|
8.297107696533203,
|
|
7.952710151672363,
|
|
7.728457927703857,
|
|
9.069082260131836,
|
|
6.80143404006958,
|
|
6.168771743774414,
|
|
7.780761241912842,
|
|
7.264509677886963,
|
|
7.721634387969971,
|
|
5.931019306182861,
|
|
8.71249771118164,
|
|
7.045263290405273,
|
|
5.595153331756592,
|
|
8.606344223022461,
|
|
7.333461284637451,
|
|
7.434794902801514,
|
|
4.909368515014648,
|
|
6.529274940490723,
|
|
3.2044527530670166,
|
|
4.450833320617676,
|
|
9.15864086151123,
|
|
7.603370189666748,
|
|
7.163464069366455,
|
|
4.514288902282715,
|
|
2.936744213104248,
|
|
6.017610549926758,
|
|
6.448644161224365,
|
|
8.636395454406738,
|
|
6.373209476470947,
|
|
7.272717475891113,
|
|
4.8009185791015625,
|
|
6.993277072906494,
|
|
7.068300724029541,
|
|
7.53340482711792,
|
|
7.4401326179504395,
|
|
7.977913856506348,
|
|
9.181097030639648,
|
|
7.183773994445801,
|
|
6.776640892028809,
|
|
6.810145378112793,
|
|
6.086609840393066,
|
|
9.078044891357422,
|
|
5.633232593536377,
|
|
7.695226669311523,
|
|
5.442765712738037,
|
|
8.75350284576416,
|
|
7.758969783782959,
|
|
7.245949745178223,
|
|
7.80985164642334,
|
|
6.605112075805664,
|
|
7.24437952041626,
|
|
7.7778215408325195,
|
|
8.456467628479004,
|
|
5.285576343536377,
|
|
8.28867244720459,
|
|
7.879434585571289,
|
|
8.340057373046875,
|
|
5.838737487792969,
|
|
8.670787811279297,
|
|
8.561763763427734,
|
|
8.80904769897461,
|
|
5.523489952087402,
|
|
8.205552101135254,
|
|
5.81448221206665,
|
|
6.502568244934082,
|
|
5.51200532913208,
|
|
6.332709789276123,
|
|
5.85950231552124,
|
|
7.721321105957031,
|
|
7.371209144592285,
|
|
5.3772382736206055,
|
|
7.831151962280273,
|
|
6.771039009094238,
|
|
5.647019863128662,
|
|
3.3475260734558105,
|
|
9.21485710144043,
|
|
6.554588317871094,
|
|
7.803776741027832,
|
|
5.230503559112549,
|
|
7.31123685836792,
|
|
7.461449146270752,
|
|
5.785803318023682,
|
|
2.818866491317749,
|
|
7.119564533233643,
|
|
7.815005779266357,
|
|
7.14105749130249,
|
|
7.022451400756836,
|
|
8.005674362182617,
|
|
7.6263227462768555,
|
|
7.574337482452393,
|
|
6.168295383453369,
|
|
6.522130012512207,
|
|
8.820441246032715,
|
|
8.641220092773438,
|
|
8.199234008789062,
|
|
4.685672760009766,
|
|
6.580758571624756,
|
|
6.7318220138549805,
|
|
7.216886043548584,
|
|
4.987853050231934,
|
|
6.9638471603393555,
|
|
8.238450050354004,
|
|
6.355881690979004,
|
|
8.457653045654297,
|
|
8.574877738952637,
|
|
8.558584213256836,
|
|
8.179498672485352,
|
|
8.395395278930664,
|
|
5.779758453369141,
|
|
5.897271633148193,
|
|
5.965787410736084,
|
|
7.879891872406006,
|
|
7.1940083503723145,
|
|
7.250895023345947,
|
|
7.340498447418213,
|
|
7.3146209716796875,
|
|
7.630643367767334,
|
|
5.256970405578613,
|
|
6.986878871917725,
|
|
5.032907962799072,
|
|
6.915760040283203,
|
|
7.389677047729492,
|
|
7.766031265258789,
|
|
7.362154483795166,
|
|
7.522637844085693,
|
|
4.709517955780029,
|
|
6.954688549041748,
|
|
6.788074493408203,
|
|
7.9603118896484375,
|
|
8.153197288513184,
|
|
7.945971488952637,
|
|
5.763076305389404,
|
|
8.035938262939453,
|
|
7.177386283874512,
|
|
7.629238128662109,
|
|
8.1404390335083,
|
|
4.857499122619629,
|
|
7.7081756591796875,
|
|
7.729892730712891,
|
|
5.2494425773620605,
|
|
7.856828212738037,
|
|
7.413257122039795,
|
|
5.691137313842773,
|
|
6.185434341430664,
|
|
6.53693151473999,
|
|
8.347500801086426,
|
|
8.713299751281738,
|
|
8.910021781921387,
|
|
8.06331729888916,
|
|
8.161259651184082,
|
|
6.673550128936768,
|
|
7.395747661590576,
|
|
6.544902801513672,
|
|
7.371769428253174,
|
|
8.319907188415527,
|
|
6.7722697257995605,
|
|
7.3024749755859375,
|
|
8.515557289123535,
|
|
7.880080699920654,
|
|
10.560447692871094,
|
|
8.548553466796875,
|
|
8.010724067687988,
|
|
8.251697540283203,
|
|
9.363635063171387,
|
|
10.383763313293457,
|
|
8.954550743103027,
|
|
7.073766708374023,
|
|
7.3394365310668945,
|
|
7.901332855224609,
|
|
5.292531967163086,
|
|
7.994369983673096,
|
|
7.169919967651367,
|
|
8.937761306762695,
|
|
7.052704334259033,
|
|
7.712167739868164,
|
|
6.639589786529541,
|
|
4.640880584716797,
|
|
6.953775882720947,
|
|
7.011972427368164,
|
|
6.708223342895508,
|
|
7.55882453918457,
|
|
7.379924774169922,
|
|
7.388876438140869,
|
|
6.607176303863525,
|
|
6.295664310455322,
|
|
6.873457431793213,
|
|
6.685941219329834,
|
|
7.678892612457275,
|
|
6.277175426483154,
|
|
6.82502555847168,
|
|
6.493975639343262,
|
|
5.599217414855957,
|
|
2.995514392852783,
|
|
4.2061686515808105,
|
|
5.388845443725586,
|
|
6.046504497528076,
|
|
6.199982643127441,
|
|
7.248841285705566,
|
|
6.691074848175049,
|
|
5.309595108032227,
|
|
2.932786226272583,
|
|
2.7796411514282227,
|
|
6.531428813934326,
|
|
3.3787026405334473,
|
|
6.607399940490723,
|
|
5.987377643585205,
|
|
5.107828617095947,
|
|
6.891719818115234,
|
|
6.07973575592041,
|
|
5.89137077331543,
|
|
4.002294540405273,
|
|
4.991847991943359,
|
|
6.058988571166992,
|
|
6.652078628540039,
|
|
5.368412017822266,
|
|
6.383002758026123,
|
|
5.716227054595947,
|
|
5.6958794593811035,
|
|
5.975515842437744,
|
|
6.719594955444336,
|
|
2.4692540168762207,
|
|
2.7202696800231934,
|
|
2.5226945877075195,
|
|
2.80049729347229,
|
|
2.9589719772338867,
|
|
2.723951816558838,
|
|
2.7555041313171387,
|
|
2.871811866760254,
|
|
2.7948708534240723,
|
|
2.851465940475464,
|
|
2.599896192550659,
|
|
2.8908488750457764,
|
|
2.9524176120758057
|
|
]
|
|
},
|
|
"verification": {
|
|
"all_ternary": true,
|
|
"failed_layers": []
|
|
},
|
|
"perplexity": 3012.731150040198
|
|
} |