852 lines
26 KiB
JSON
852 lines
26 KiB
JSON
|
{
|
||
|
|
"prompt": "A playful orange tabby cat starts in an ordinary cozy living room in a normal house, afternoon light, sofa and rug. The cat crouches, jumps, and does one clean athletic backflip in slow motion. As the backflip completes there is a sharp cinematic cut: the cat lands perfectly on a glowing neon disco dance floor wearing oversized black sunglasses. Mirror ball reflections, colorful lights, joyful party energy, stylish and funny, clear before-and-after transformation.",
|
||
|
|
"model_path": "/models/minimax_h3_fl2va_pruned_nvfp4.safetensors",
|
||
|
|
"width": 1344,
|
||
|
|
"height": 768,
|
||
|
|
"frames": 124,
|
||
|
|
"steps": 12,
|
||
|
|
"sampler_step": 1,
|
||
|
|
"seed": 440420,
|
||
|
|
"text_tokens": 100,
|
||
|
|
"block_index": 24,
|
||
|
|
"attention": "sage2",
|
||
|
|
"warmup": 2,
|
||
|
|
"iterations": 3,
|
||
|
|
"hidden_shape": [
|
||
|
|
37810,
|
||
|
|
5376
|
||
|
|
],
|
||
|
|
"output_shape": [
|
||
|
|
37810,
|
||
|
|
5376
|
||
|
|
],
|
||
|
|
"segments": [
|
||
|
|
[
|
||
|
|
0,
|
||
|
|
100,
|
||
|
|
1
|
||
|
|
],
|
||
|
|
[
|
||
|
|
100,
|
||
|
|
514,
|
||
|
|
2
|
||
|
|
],
|
||
|
|
[
|
||
|
|
514,
|
||
|
|
37810,
|
||
|
|
0
|
||
|
|
]
|
||
|
|
],
|
||
|
|
"timings": {
|
||
|
|
"norm1": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0036721593351103365,
|
||
|
|
"p50_s": 0.0036359680234454572,
|
||
|
|
"p90_s": 0.00376709041884169,
|
||
|
|
"p95_s": 0.003783480718266219,
|
||
|
|
"p99_s": 0.003796592957805842,
|
||
|
|
"min_s": 0.0035806389641948044,
|
||
|
|
"max_s": 0.003799871017690748
|
||
|
|
},
|
||
|
|
"modulate_msa": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.010726557637099177,
|
||
|
|
"p50_s": 0.010602172988001257,
|
||
|
|
"p90_s": 0.010938596178311855,
|
||
|
|
"p95_s": 0.010980649077100679,
|
||
|
|
"p99_s": 0.01101429139613174,
|
||
|
|
"min_s": 0.010554797947406769,
|
||
|
|
"max_s": 0.011022701975889504
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.flatten_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 5.6746648624539375e-06,
|
||
|
|
"p50_s": 4.847999662160873e-06,
|
||
|
|
"p90_s": 6.84480182826519e-06,
|
||
|
|
"p95_s": 7.094402099028229e-06,
|
||
|
|
"p99_s": 7.294082315638661e-06,
|
||
|
|
"min_s": 4.83199255540967e-06,
|
||
|
|
"max_s": 7.344002369791269e-06
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.pre_quant_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0,
|
||
|
|
"p50_s": 0.0,
|
||
|
|
"p90_s": 0.0,
|
||
|
|
"p95_s": 0.0,
|
||
|
|
"p99_s": 0.0,
|
||
|
|
"min_s": 0.0,
|
||
|
|
"max_s": 0.0
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.packed_weight_wrapper": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 1.602132882302006e-05,
|
||
|
|
"p50_s": 1.003197394311428e-05,
|
||
|
|
"p90_s": 2.5776028633117675e-05,
|
||
|
|
"p95_s": 2.7744035469368098e-05,
|
||
|
|
"p99_s": 2.9318440938368442e-05,
|
||
|
|
"min_s": 8.319970220327377e-06,
|
||
|
|
"max_s": 2.9712042305618525e-05
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.bias_cast": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 3.562658093869686e-06,
|
||
|
|
"p50_s": 3.5199918784201145e-06,
|
||
|
|
"p90_s": 3.68637265637517e-06,
|
||
|
|
"p95_s": 3.7071702536195515e-06,
|
||
|
|
"p99_s": 3.7238083314150572e-06,
|
||
|
|
"min_s": 3.4400145523250103e-06,
|
||
|
|
"max_s": 3.7279678508639336e-06
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.activation_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.001872149354312569,
|
||
|
|
"p50_s": 0.0017504799761809409,
|
||
|
|
"p90_s": 0.002067510422784835,
|
||
|
|
"p95_s": 0.0021071392286103217,
|
||
|
|
"p99_s": 0.0021388422732707114,
|
||
|
|
"min_s": 0.0017192000523209572,
|
||
|
|
"max_s": 0.0021467680344358087
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.scale_to_device": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 1.5653010147313278e-05,
|
||
|
|
"p50_s": 1.8463993910700083e-05,
|
||
|
|
"p90_s": 1.87960104085505e-05,
|
||
|
|
"p95_s": 1.8837512470781803e-05,
|
||
|
|
"p99_s": 1.8870714120566843e-05,
|
||
|
|
"min_s": 9.616021998226643e-06,
|
||
|
|
"max_s": 1.8879014533013105e-05
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.activation_quant_pack": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.002274436662749698,
|
||
|
|
"p50_s": 0.0022630870225839317,
|
||
|
|
"p90_s": 0.002373551798518747,
|
||
|
|
"p95_s": 0.002387359895510599,
|
||
|
|
"p99_s": 0.0023984063731040807,
|
||
|
|
"min_s": 0.0021590549731627107,
|
||
|
|
"max_s": 0.002401167992502451
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.activation_quant_wrap": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 6.943999323993921e-06,
|
||
|
|
"p50_s": 6.480026058852673e-06,
|
||
|
|
"p90_s": 7.785600610077381e-06,
|
||
|
|
"p95_s": 7.948797428980469e-06,
|
||
|
|
"p99_s": 8.07935488410294e-06,
|
||
|
|
"min_s": 6.239977665245533e-06,
|
||
|
|
"max_s": 8.111994247883558e-06
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.gemm": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.02569908833053584,
|
||
|
|
"p50_s": 0.025752091023605317,
|
||
|
|
"p90_s": 0.02613150858087465,
|
||
|
|
"p95_s": 0.02617893577553332,
|
||
|
|
"p99_s": 0.026216877531260254,
|
||
|
|
"min_s": 0.025118810997810215,
|
||
|
|
"max_s": 0.026226362970191985
|
||
|
|
},
|
||
|
|
"linear.attn_qkv_proj.slice_reshape": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 6.245332770049572e-06,
|
||
|
|
"p50_s": 5.568028427660465e-06,
|
||
|
|
"p90_s": 7.244781590998172e-06,
|
||
|
|
"p95_s": 7.454375736415386e-06,
|
||
|
|
"p99_s": 7.622051052749157e-06,
|
||
|
|
"min_s": 5.504000000655651e-06,
|
||
|
|
"max_s": 7.6639698818326e-06
|
||
|
|
},
|
||
|
|
"attn_qkv_proj": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.029945935336096834,
|
||
|
|
"p50_s": 0.02986503398278728,
|
||
|
|
"p90_s": 0.030461424414534123,
|
||
|
|
"p95_s": 0.030535973218502478,
|
||
|
|
"p99_s": 0.03059561226167716,
|
||
|
|
"min_s": 0.029362250003032386,
|
||
|
|
"max_s": 0.030610522022470832
|
||
|
|
},
|
||
|
|
"attn_qkv_split_view": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 1.9002686409900587e-05,
|
||
|
|
"p50_s": 2.0368024706840515e-05,
|
||
|
|
"p90_s": 2.22496222704649e-05,
|
||
|
|
"p95_s": 2.2484821965917948e-05,
|
||
|
|
"p99_s": 2.2672981722280386e-05,
|
||
|
|
"min_s": 1.392001286149025e-05,
|
||
|
|
"max_s": 2.2720021661370993e-05
|
||
|
|
},
|
||
|
|
"attn_qk_rms_rope": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.012115544329086939,
|
||
|
|
"p50_s": 0.012150077964179218,
|
||
|
|
"p90_s": 0.012263204413466157,
|
||
|
|
"p95_s": 0.012277345219627023,
|
||
|
|
"p99_s": 0.012288657864555717,
|
||
|
|
"min_s": 0.01190506899729371,
|
||
|
|
"max_s": 0.01229148602578789
|
||
|
|
},
|
||
|
|
"attn_q_transpose_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.005235358665231615,
|
||
|
|
"p50_s": 0.005258975026663393,
|
||
|
|
"p90_s": 0.005414558190386742,
|
||
|
|
"p95_s": 0.005434006085852161,
|
||
|
|
"p99_s": 0.005449564402224496,
|
||
|
|
"min_s": 0.004993646987713873,
|
||
|
|
"max_s": 0.00545345398131758
|
||
|
|
},
|
||
|
|
"attn_k_transpose_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.005438665684778243,
|
||
|
|
"p50_s": 0.00541571102803573,
|
||
|
|
"p90_s": 0.005487903009634465,
|
||
|
|
"p95_s": 0.005496927007334306,
|
||
|
|
"p99_s": 0.00550414620549418,
|
||
|
|
"min_s": 0.005394335021264851,
|
||
|
|
"max_s": 0.005505951005034149
|
||
|
|
},
|
||
|
|
"attn_v_transpose_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0051273749947237475,
|
||
|
|
"p50_s": 0.005076878995168954,
|
||
|
|
"p90_s": 0.0052603541989810765,
|
||
|
|
"p95_s": 0.005283288599457592,
|
||
|
|
"p99_s": 0.005301636119838804,
|
||
|
|
"min_s": 0.00499902298906818,
|
||
|
|
"max_s": 0.005306222999934107
|
||
|
|
},
|
||
|
|
"attention_kernel": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.2554921513268103,
|
||
|
|
"p50_s": 0.255616124952212,
|
||
|
|
"p90_s": 0.25568432337604463,
|
||
|
|
"p95_s": 0.2556928481790237,
|
||
|
|
"p99_s": 0.255699668021407,
|
||
|
|
"min_s": 0.2551589560462162,
|
||
|
|
"max_s": 0.2557013729820028
|
||
|
|
},
|
||
|
|
"attn_output_reshape": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.004698617655473451,
|
||
|
|
"p50_s": 0.0046985429944470525,
|
||
|
|
"p90_s": 0.004699272592552007,
|
||
|
|
"p95_s": 0.004699363792315126,
|
||
|
|
"p99_s": 0.004699436752125621,
|
||
|
|
"min_s": 0.004697854979895055,
|
||
|
|
"max_s": 0.004699454992078245
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.flatten_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 4.762647828708093e-06,
|
||
|
|
"p50_s": 4.624016582965851e-06,
|
||
|
|
"p90_s": 5.2511692047119146e-06,
|
||
|
|
"p95_s": 5.329563282430173e-06,
|
||
|
|
"p99_s": 5.392278544604778e-06,
|
||
|
|
"min_s": 4.255969543009996e-06,
|
||
|
|
"max_s": 5.40795736014843e-06
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.pre_quant_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0,
|
||
|
|
"p50_s": 0.0,
|
||
|
|
"p90_s": 0.0,
|
||
|
|
"p95_s": 0.0,
|
||
|
|
"p99_s": 0.0,
|
||
|
|
"min_s": 0.0,
|
||
|
|
"max_s": 0.0
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.packed_weight_wrapper": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 1.0826683137565851e-05,
|
||
|
|
"p50_s": 1.0176037903875113e-05,
|
||
|
|
"p90_s": 1.1814420577138663e-05,
|
||
|
|
"p95_s": 1.2019218411296606e-05,
|
||
|
|
"p99_s": 1.218305667862296e-05,
|
||
|
|
"min_s": 1.0079995263367891e-05,
|
||
|
|
"max_s": 1.222401624545455e-05
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.bias_cast": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 3.573348900924126e-06,
|
||
|
|
"p50_s": 3.5840203054249287e-06,
|
||
|
|
"p90_s": 3.6223907954990866e-06,
|
||
|
|
"p95_s": 3.627187106758356e-06,
|
||
|
|
"p99_s": 3.6310241557657717e-06,
|
||
|
|
"min_s": 3.5040429793298244e-06,
|
||
|
|
"max_s": 3.631983418017626e-06
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.activation_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.002271252994736036,
|
||
|
|
"p50_s": 0.002267103991471231,
|
||
|
|
"p90_s": 0.002277253591455519,
|
||
|
|
"p95_s": 0.002278522291453555,
|
||
|
|
"p99_s": 0.002279537251451984,
|
||
|
|
"min_s": 0.0022668640012852848,
|
||
|
|
"max_s": 0.0022797909914515913
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.scale_to_device": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 1.067200597996513e-05,
|
||
|
|
"p50_s": 1.0624004062265158e-05,
|
||
|
|
"p90_s": 1.105921110138297e-05,
|
||
|
|
"p95_s": 1.1113611981272697e-05,
|
||
|
|
"p99_s": 1.1157132685184478e-05,
|
||
|
|
"min_s": 1.022400101646781e-05,
|
||
|
|
"max_s": 1.1168012861162424e-05
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.activation_quant_pack": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0028463726630434394,
|
||
|
|
"p50_s": 0.0028423359617590904,
|
||
|
|
"p90_s": 0.00285377842374146,
|
||
|
|
"p95_s": 0.002855208731489256,
|
||
|
|
"p99_s": 0.002856352977687493,
|
||
|
|
"min_s": 0.0028401429881341755,
|
||
|
|
"max_s": 0.002856639039237052
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.activation_quant_wrap": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 6.75199165319403e-06,
|
||
|
|
"p50_s": 6.671994924545288e-06,
|
||
|
|
"p90_s": 7.1072019636631005e-06,
|
||
|
|
"p95_s": 7.1616028435528275e-06,
|
||
|
|
"p99_s": 7.205123547464608e-06,
|
||
|
|
"min_s": 6.367976311594248e-06,
|
||
|
|
"max_s": 7.2160037234425545e-06
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.gemm": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.00781480900089567,
|
||
|
|
"p50_s": 0.007808863010723144,
|
||
|
|
"p90_s": 0.007827895798254759,
|
||
|
|
"p95_s": 0.00783027489669621,
|
||
|
|
"p99_s": 0.00783217817544937,
|
||
|
|
"min_s": 0.007802909996826202,
|
||
|
|
"max_s": 0.007832653995137662
|
||
|
|
},
|
||
|
|
"linear.attn_out_proj.slice_reshape": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 5.680000564704339e-06,
|
||
|
|
"p50_s": 5.664012860506773e-06,
|
||
|
|
"p90_s": 5.791976582258939e-06,
|
||
|
|
"p95_s": 5.80797204747796e-06,
|
||
|
|
"p99_s": 5.820768419653177e-06,
|
||
|
|
"min_s": 5.552021320909262e-06,
|
||
|
|
"max_s": 5.8239675126969814e-06
|
||
|
|
},
|
||
|
|
"attn_out_proj": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.013016370648983866,
|
||
|
|
"p50_s": 0.013008972979150712,
|
||
|
|
"p90_s": 0.013036953774280846,
|
||
|
|
"p95_s": 0.013040451373672112,
|
||
|
|
"p99_s": 0.013043249453185126,
|
||
|
|
"min_s": 0.012996189994737506,
|
||
|
|
"max_s": 0.01304394897306338
|
||
|
|
},
|
||
|
|
"gate_msa": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.008527667979554584,
|
||
|
|
"p50_s": 0.008532046980690211,
|
||
|
|
"p90_s": 0.00853673176607117,
|
||
|
|
"p95_s": 0.00853731736424379,
|
||
|
|
"p99_s": 0.008537785842781887,
|
||
|
|
"min_s": 0.00851305399555713,
|
||
|
|
"max_s": 0.00853790296241641
|
||
|
|
},
|
||
|
|
"norm2": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0035164790072788796,
|
||
|
|
"p50_s": 0.0035144950379617512,
|
||
|
|
"p90_s": 0.003523019806016237,
|
||
|
|
"p95_s": 0.0035240854020230474,
|
||
|
|
"p99_s": 0.0035249378788284958,
|
||
|
|
"min_s": 0.0035097909858450294,
|
||
|
|
"max_s": 0.003525150998029858
|
||
|
|
},
|
||
|
|
"modulate_mlp": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.010573251328120628,
|
||
|
|
"p50_s": 0.010518733994103968,
|
||
|
|
"p90_s": 0.010858023609034716,
|
||
|
|
"p95_s": 0.01090043481090106,
|
||
|
|
"p99_s": 0.010934363772394136,
|
||
|
|
"min_s": 0.010258173977490515,
|
||
|
|
"max_s": 0.010942846012767404
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.flatten_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 4.634668584913015e-06,
|
||
|
|
"p50_s": 4.608009476214647e-06,
|
||
|
|
"p90_s": 5.004799459129572e-06,
|
||
|
|
"p95_s": 5.054398206993937e-06,
|
||
|
|
"p99_s": 5.09407720528543e-06,
|
||
|
|
"min_s": 4.191999323666096e-06,
|
||
|
|
"max_s": 5.103996954858303e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.pre_quant_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0,
|
||
|
|
"p50_s": 0.0,
|
||
|
|
"p90_s": 0.0,
|
||
|
|
"p95_s": 0.0,
|
||
|
|
"p99_s": 0.0,
|
||
|
|
"min_s": 0.0,
|
||
|
|
"max_s": 0.0
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.packed_weight_wrapper": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 8.896032037834326e-06,
|
||
|
|
"p50_s": 8.848030120134354e-06,
|
||
|
|
"p90_s": 9.232014417648316e-06,
|
||
|
|
"p95_s": 9.28001245483756e-06,
|
||
|
|
"p99_s": 9.318410884588956e-06,
|
||
|
|
"min_s": 8.51205550134182e-06,
|
||
|
|
"max_s": 9.328010492026806e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.bias_cast": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 3.4399951497713723e-06,
|
||
|
|
"p50_s": 3.439956344664097e-06,
|
||
|
|
"p90_s": 3.516790457069874e-06,
|
||
|
|
"p95_s": 3.526394721120596e-06,
|
||
|
|
"p99_s": 3.5340781323611736e-06,
|
||
|
|
"min_s": 3.3440301194787025e-06,
|
||
|
|
"max_s": 3.535998985171318e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.activation_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0017615193501114845,
|
||
|
|
"p50_s": 0.0017409750143997371,
|
||
|
|
"p90_s": 0.0018063198192976415,
|
||
|
|
"p95_s": 0.0018144879199098797,
|
||
|
|
"p99_s": 0.00182102240039967,
|
||
|
|
"min_s": 0.0017209270154125988,
|
||
|
|
"max_s": 0.0018226560205221176
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.scale_to_device": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 9.466661140322685e-06,
|
||
|
|
"p50_s": 9.503972250968218e-06,
|
||
|
|
"p90_s": 9.542389307171106e-06,
|
||
|
|
"p95_s": 9.547191439196467e-06,
|
||
|
|
"p99_s": 9.551033144816756e-06,
|
||
|
|
"min_s": 9.34401759877801e-06,
|
||
|
|
"max_s": 9.551993571221828e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.activation_quant_pack": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.002172234351746738,
|
||
|
|
"p50_s": 0.0021623510401695967,
|
||
|
|
"p90_s": 0.0021976286079734564,
|
||
|
|
"p95_s": 0.002202038303948939,
|
||
|
|
"p99_s": 0.002205566060729325,
|
||
|
|
"min_s": 0.002147904015146196,
|
||
|
|
"max_s": 0.0022064479999244213
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.activation_quant_wrap": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 6.416017034401496e-06,
|
||
|
|
"p50_s": 6.399990525096655e-06,
|
||
|
|
"p90_s": 6.489630322903395e-06,
|
||
|
|
"p95_s": 6.500835297629237e-06,
|
||
|
|
"p99_s": 6.509799277409911e-06,
|
||
|
|
"min_s": 6.336020305752754e-06,
|
||
|
|
"max_s": 6.51204027235508e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.gemm": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.07115556466548394,
|
||
|
|
"p50_s": 0.07123408204643056,
|
||
|
|
"p90_s": 0.07148396357661113,
|
||
|
|
"p95_s": 0.0715151987678837,
|
||
|
|
"p99_s": 0.07154018692090176,
|
||
|
|
"min_s": 0.07068617799086496,
|
||
|
|
"max_s": 0.07154643395915627
|
||
|
|
},
|
||
|
|
"linear.mlp_fc1.slice_reshape": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 5.6160109428068e-06,
|
||
|
|
"p50_s": 5.664012860506773e-06,
|
||
|
|
"p90_s": 5.715189035981893e-06,
|
||
|
|
"p95_s": 5.721586057916284e-06,
|
||
|
|
"p99_s": 5.726703675463796e-06,
|
||
|
|
"min_s": 5.456036888062954e-06,
|
||
|
|
"max_s": 5.727983079850674e-06
|
||
|
|
},
|
||
|
|
"mlp_fc1": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.07517527565748121,
|
||
|
|
"p50_s": 0.07530217699240893,
|
||
|
|
"p90_s": 0.07546080739703029,
|
||
|
|
"p95_s": 0.07548063619760798,
|
||
|
|
"p99_s": 0.0754964992380701,
|
||
|
|
"min_s": 0.07472318498184904,
|
||
|
|
"max_s": 0.07550046499818563
|
||
|
|
},
|
||
|
|
"mlp_swiglu": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.025796906343506027,
|
||
|
|
"p50_s": 0.025601482018828392,
|
||
|
|
"p90_s": 0.026077552419155835,
|
||
|
|
"p95_s": 0.026137061219196768,
|
||
|
|
"p99_s": 0.02618466825922951,
|
||
|
|
"min_s": 0.025592666992451996,
|
||
|
|
"max_s": 0.026196570019237697
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.flatten_contiguous": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 4.560026961068313e-06,
|
||
|
|
"p50_s": 4.56004636362195e-06,
|
||
|
|
"p90_s": 4.726427141577005e-06,
|
||
|
|
"p95_s": 4.747224738821387e-06,
|
||
|
|
"p99_s": 4.763862816616893e-06,
|
||
|
|
"min_s": 4.352012183517218e-06,
|
||
|
|
"max_s": 4.768022336065769e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.pre_quant_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0,
|
||
|
|
"p50_s": 0.0,
|
||
|
|
"p90_s": 0.0,
|
||
|
|
"p95_s": 0.0,
|
||
|
|
"p99_s": 0.0,
|
||
|
|
"min_s": 0.0,
|
||
|
|
"max_s": 0.0
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.packed_weight_wrapper": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 8.975989961375793e-06,
|
||
|
|
"p50_s": 9.119976311922073e-06,
|
||
|
|
"p90_s": 9.299209341406823e-06,
|
||
|
|
"p95_s": 9.321613470092414e-06,
|
||
|
|
"p99_s": 9.33953677304089e-06,
|
||
|
|
"min_s": 8.463975973427296e-06,
|
||
|
|
"max_s": 9.34401759877801e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.bias_cast": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 3.5733294983704886e-06,
|
||
|
|
"p50_s": 3.5520060919225216e-06,
|
||
|
|
"p90_s": 3.6544050090014937e-06,
|
||
|
|
"p95_s": 3.667204873636365e-06,
|
||
|
|
"p99_s": 3.677444765344262e-06,
|
||
|
|
"min_s": 3.4879776649177074e-06,
|
||
|
|
"max_s": 3.6800047382712364e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.activation_scale": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.00451310898642987,
|
||
|
|
"p50_s": 0.004505342978518456,
|
||
|
|
"p90_s": 0.004527206195052713,
|
||
|
|
"p95_s": 0.004529939097119496,
|
||
|
|
"p99_s": 0.004532125418772921,
|
||
|
|
"min_s": 0.004501311981584877,
|
||
|
|
"max_s": 0.004532671999186277
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.scale_to_device": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 9.279989171773195e-06,
|
||
|
|
"p50_s": 9.535986464470625e-06,
|
||
|
|
"p90_s": 9.753589984029531e-06,
|
||
|
|
"p95_s": 9.780790423974394e-06,
|
||
|
|
"p99_s": 9.802550775930286e-06,
|
||
|
|
"min_s": 8.495990186929703e-06,
|
||
|
|
"max_s": 9.807990863919258e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.activation_quant_pack": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.005867129017133266,
|
||
|
|
"p50_s": 0.005761470005381852,
|
||
|
|
"p90_s": 0.0060237164259888225,
|
||
|
|
"p95_s": 0.006056497228564694,
|
||
|
|
"p99_s": 0.006082721870625391,
|
||
|
|
"min_s": 0.005750639014877379,
|
||
|
|
"max_s": 0.006089278031140566
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.activation_quant_wrap": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 6.517337169498205e-06,
|
||
|
|
"p50_s": 6.52798917144537e-06,
|
||
|
|
"p90_s": 6.681610830128194e-06,
|
||
|
|
"p95_s": 6.700813537463546e-06,
|
||
|
|
"p99_s": 6.716175703331828e-06,
|
||
|
|
"min_s": 6.304006092250347e-06,
|
||
|
|
"max_s": 6.720016244798899e-06
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.gemm": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.0607306759726877,
|
||
|
|
"p50_s": 0.058016403985675424,
|
||
|
|
"p90_s": 0.06506514557404444,
|
||
|
|
"p95_s": 0.06594623827259057,
|
||
|
|
"p99_s": 0.06665111243142748,
|
||
|
|
"min_s": 0.05734829296125099,
|
||
|
|
"max_s": 0.06682733097113669
|
||
|
|
},
|
||
|
|
"linear.mlp_fc2.slice_reshape": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 5.6586771582563715e-06,
|
||
|
|
"p50_s": 5.567970219999552e-06,
|
||
|
|
"p90_s": 5.849602166563272e-06,
|
||
|
|
"p95_s": 5.884806159883737e-06,
|
||
|
|
"p99_s": 5.912969354540109e-06,
|
||
|
|
"min_s": 5.488051101565361e-06,
|
||
|
|
"max_s": 5.920010153204203e-06
|
||
|
|
},
|
||
|
|
"mlp_fc2": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.07119058868071686,
|
||
|
|
"p50_s": 0.06835102604236454,
|
||
|
|
"p90_s": 0.07540559160988779,
|
||
|
|
"p95_s": 0.07628741230582818,
|
||
|
|
"p99_s": 0.07699286886258051,
|
||
|
|
"min_s": 0.06805150699801743,
|
||
|
|
"max_s": 0.07716923300176859
|
||
|
|
},
|
||
|
|
"gate_mlp": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.013778018668138733,
|
||
|
|
"p50_s": 0.010043838003184646,
|
||
|
|
"p90_s": 0.02003887001192197,
|
||
|
|
"p95_s": 0.021288249013014137,
|
||
|
|
"p99_s": 0.02228775221388787,
|
||
|
|
"min_s": 0.008752589987125248,
|
||
|
|
"max_s": 0.022537628014106303
|
||
|
|
},
|
||
|
|
"block_total": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.5541369766773035,
|
||
|
|
"p50_s": 0.5570170399732888,
|
||
|
|
"p90_s": 0.5594101664144546,
|
||
|
|
"p95_s": 0.5597093072196002,
|
||
|
|
"p99_s": 0.5599486198637169,
|
||
|
|
"min_s": 0.5453854420338757,
|
||
|
|
"max_s": 0.560008448024746
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"module_forward": {
|
||
|
|
"count": 3,
|
||
|
|
"mean_s": 0.5096157593264555,
|
||
|
|
"p50_s": 0.5090722020249814,
|
||
|
|
"p90_s": 0.5104809443932027,
|
||
|
|
"p95_s": 0.5106570371892303,
|
||
|
|
"p99_s": 0.5107979114260524,
|
||
|
|
"min_s": 0.5089419459691271,
|
||
|
|
"max_s": 0.510833129985258
|
||
|
|
},
|
||
|
|
"module_forward_checksum": 296883328.0,
|
||
|
|
"fused_elementwise": true,
|
||
|
|
"profiler_summary": {
|
||
|
|
"profiled_iterations": 1,
|
||
|
|
"runtime_kernel_launches": 0,
|
||
|
|
"runtime_kernel_launches_per_block": 0.0,
|
||
|
|
"positive_self_device_allocated_bytes": 9093546496
|
||
|
|
},
|
||
|
|
"profiler_top_events": [
|
||
|
|
{
|
||
|
|
"key": "aten::rms_norm",
|
||
|
|
"count": 2,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 1367.4550000000002,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": -303104,
|
||
|
|
"input_shapes": "[[37810, 5376], [], [5376], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::_fused_rms_norm",
|
||
|
|
"count": 2,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 1346.1589999999997,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 5376], [], [5376], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::quantize_nvfp4",
|
||
|
|
"count": 2,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 1125.328,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 5376], [], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::zeros",
|
||
|
|
"count": 4,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 666.7999999999993,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[], [], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::zero_",
|
||
|
|
"count": 2,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 591.2480000000005,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37888, 336]]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::linear",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 588.7200000000003,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 5376], [21504, 5376], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 445.0400000000004,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": -512,
|
||
|
|
"input_shapes": "[[37824, 2688], [21504, 2688], [], [], [37888, 336], [21504, 336], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::rms_rope_split_half_",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 305.08799999999974,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[1, 37810, 56, 128], [1, 37810, 56, 128], [1, 37810, 1, 48, 2, 2], [128], [128], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::_to_copy",
|
||
|
|
"count": 8,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 261.40800000000036,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[], [], [], [], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::linear",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 210.1280000000006,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 7168], [5376, 7168], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::to",
|
||
|
|
"count": 12,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 192.3839999999982,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[], [], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::copy_",
|
||
|
|
"count": 8,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 184.64000000000033,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 173.08799999999974,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": -512,
|
||
|
|
"input_shapes": "[[37824, 3584], [5376, 3584], [], [], [37888, 448], [5376, 448], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::linear",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 154.47999999999956,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 14336], [5376, 14336], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 136.70399999999972,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": -512,
|
||
|
|
"input_shapes": "[[37824, 7168], [5376, 7168], [], [], [37888, 896], [5376, 896], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::empty",
|
||
|
|
"count": 34,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 132.73599999999925,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 8009298432,
|
||
|
|
"input_shapes": "[[], [], [], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::linear",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 129.66399999999976,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 0,
|
||
|
|
"input_shapes": "[[37810, 5376], [28672, 5376], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 109.07200000000012,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": -512,
|
||
|
|
"input_shapes": "[[37824, 2688], [28672, 2688], [], [], [37888, 336], [28672, 336], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::mean",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 99.28000000000065,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 14336,
|
||
|
|
"input_shapes": "[[1, 37810, 56, 128], [], [], []]"
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"key": "aten::silu",
|
||
|
|
"count": 1,
|
||
|
|
"device_type": "DeviceType.CPU",
|
||
|
|
"cpu_time_total_us": 99.15200000000004,
|
||
|
|
"device_time_total_us": 0.0,
|
||
|
|
"self_device_time_total_us": 0.0,
|
||
|
|
"self_device_memory_usage_bytes": 1084227584,
|
||
|
|
"input_shapes": "[[37810, 14336]]"
|
||
|
|
}
|
||
|
|
]
|
||
|
|
}
|