852 lines
No EOL
26 KiB
JSON
852 lines
No EOL
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": 49,
|
|
"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": 0.0036217596692343554,
|
|
"p50_s": 0.00359812198439613,
|
|
"p90_s": 0.0037101987865753474,
|
|
"p95_s": 0.0037242083868477493,
|
|
"p99_s": 0.0037354160670656712,
|
|
"min_s": 0.0035289390361867845,
|
|
"max_s": 0.0037382179871201515
|
|
},
|
|
"modulate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.01067996733278657,
|
|
"p50_s": 0.010696462995838374,
|
|
"p90_s": 0.011010011786129327,
|
|
"p95_s": 0.011049205384915696,
|
|
"p99_s": 0.011080560263944791,
|
|
"min_s": 0.010255040018819273,
|
|
"max_s": 0.011088398983702064
|
|
},
|
|
"linear.attn_qkv_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 5.525342809657256e-06,
|
|
"p50_s": 4.208006430417299e-06,
|
|
"p90_s": 7.382419425994158e-06,
|
|
"p95_s": 7.779221050441265e-06,
|
|
"p99_s": 8.096662349998951e-06,
|
|
"min_s": 4.191999323666096e-06,
|
|
"max_s": 8.176022674888372e-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": 1.5418666104475658e-05,
|
|
"p50_s": 8.544011507183313e-06,
|
|
"p90_s": 2.5260786060243845e-05,
|
|
"p95_s": 2.7350382879376408e-05,
|
|
"p99_s": 2.9022060334682465e-05,
|
|
"min_s": 8.27200710773468e-06,
|
|
"max_s": 2.9439979698508978e-05
|
|
},
|
|
"linear.attn_qkv_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.674688438574473e-06,
|
|
"p50_s": 3.6479905247688293e-06,
|
|
"p90_s": 3.827223554253578e-06,
|
|
"p95_s": 3.849627682939172e-06,
|
|
"p99_s": 3.867550985887647e-06,
|
|
"min_s": 3.5040429793298244e-06,
|
|
"max_s": 3.872031811624765e-06
|
|
},
|
|
"linear.attn_qkv_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0017329036879042785,
|
|
"p50_s": 0.0017396770417690277,
|
|
"p90_s": 0.0017431330401450396,
|
|
"p95_s": 0.001743565039942041,
|
|
"p99_s": 0.0017439106397796422,
|
|
"min_s": 0.001715036982204765,
|
|
"max_s": 0.0017439970397390425
|
|
},
|
|
"linear.attn_qkv_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.3610658546288809e-05,
|
|
"p50_s": 1.2303993571549654e-05,
|
|
"p90_s": 1.84479751624167e-05,
|
|
"p95_s": 1.9215972861275078e-05,
|
|
"p99_s": 1.983037102036178e-05,
|
|
"min_s": 8.544011507183313e-06,
|
|
"max_s": 1.9983970560133457e-05
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.0022982683343191943,
|
|
"p50_s": 0.0023595320526510477,
|
|
"p90_s": 0.0024083896074444055,
|
|
"p95_s": 0.0024144968017935755,
|
|
"p99_s": 0.002419382557272911,
|
|
"min_s": 0.0021146689541637897,
|
|
"max_s": 0.002420603996142745
|
|
},
|
|
"linear.attn_qkv_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 7.045338861644268e-06,
|
|
"p50_s": 6.511982064694166e-06,
|
|
"p90_s": 7.868779357522726e-06,
|
|
"p95_s": 8.038379019126295e-06,
|
|
"p99_s": 8.174058748409153e-06,
|
|
"min_s": 6.416055839508772e-06,
|
|
"max_s": 8.207978680729866e-06
|
|
},
|
|
"linear.attn_qkv_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.02569509532380228,
|
|
"p50_s": 0.02566098398528993,
|
|
"p90_s": 0.026279260776937007,
|
|
"p95_s": 0.026356545375892895,
|
|
"p99_s": 0.0264183730550576,
|
|
"min_s": 0.02499047201126814,
|
|
"max_s": 0.026433829974848777
|
|
},
|
|
"linear.attn_qkv_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 6.298670389999946e-06,
|
|
"p50_s": 5.504000000655651e-06,
|
|
"p90_s": 7.46243167668581e-06,
|
|
"p95_s": 7.70723563618958e-06,
|
|
"p99_s": 7.903078803792597e-06,
|
|
"min_s": 5.439971573650837e-06,
|
|
"max_s": 7.95203959569335e-06
|
|
},
|
|
"attn_qkv_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.029824565979652107,
|
|
"p50_s": 0.02990254497854039,
|
|
"p90_s": 0.030253315379377456,
|
|
"p95_s": 0.030297161679482085,
|
|
"p99_s": 0.030332238719565793,
|
|
"min_s": 0.02923014498082921,
|
|
"max_s": 0.03034100797958672
|
|
},
|
|
"attn_qkv_split_view": {
|
|
"count": 3,
|
|
"mean_s": 1.4346345172574123e-05,
|
|
"p50_s": 1.1824013199657202e-05,
|
|
"p90_s": 1.82872056029737e-05,
|
|
"p95_s": 1.909510465338826e-05,
|
|
"p99_s": 1.9741423893719913e-05,
|
|
"min_s": 1.1312018614262342e-05,
|
|
"max_s": 1.9903003703802824e-05
|
|
},
|
|
"attn_qk_rms_rope": {
|
|
"count": 3,
|
|
"mean_s": 0.012300620651027808,
|
|
"p50_s": 0.012445547967217863,
|
|
"p90_s": 0.012557216011919082,
|
|
"p95_s": 0.012571174517506735,
|
|
"p99_s": 0.012582341321976855,
|
|
"min_s": 0.011871180962771177,
|
|
"max_s": 0.012585133023094386
|
|
},
|
|
"attn_q_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.005062365322373807,
|
|
"p50_s": 0.005007288011256605,
|
|
"p90_s": 0.005225092789623886,
|
|
"p95_s": 0.005252318386919796,
|
|
"p99_s": 0.005274098864756525,
|
|
"min_s": 0.00490026397164911,
|
|
"max_s": 0.005279543984215707
|
|
},
|
|
"attn_k_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.00505693598339955,
|
|
"p50_s": 0.00496911199297756,
|
|
"p90_s": 0.005203018360771239,
|
|
"p95_s": 0.005232256656745448,
|
|
"p99_s": 0.005255647293524816,
|
|
"min_s": 0.004940201004501432,
|
|
"max_s": 0.005261494952719659
|
|
},
|
|
"attn_v_transpose_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 0.00494270701892674,
|
|
"p50_s": 0.00493679300416261,
|
|
"p90_s": 0.004955710633657872,
|
|
"p95_s": 0.004958075337344781,
|
|
"p99_s": 0.004959967100294307,
|
|
"min_s": 0.004930888011585921,
|
|
"max_s": 0.0049604400410316885
|
|
},
|
|
"attention_kernel": {
|
|
"count": 3,
|
|
"mean_s": 0.2531811236597908,
|
|
"p50_s": 0.2535298540024087,
|
|
"p90_s": 0.25376300599891694,
|
|
"p95_s": 0.25379214999848043,
|
|
"p99_s": 0.2538154651981313,
|
|
"min_s": 0.25219222297891974,
|
|
"max_s": 0.253821293998044
|
|
},
|
|
"attn_output_reshape": {
|
|
"count": 3,
|
|
"mean_s": 0.0048227393223593635,
|
|
"p50_s": 0.004690104979090393,
|
|
"p90_s": 0.005012331414036453,
|
|
"p95_s": 0.00505260971840471,
|
|
"p99_s": 0.005084832361899316,
|
|
"min_s": 0.004685224965214729,
|
|
"max_s": 0.005092888022772968
|
|
},
|
|
"linear.attn_out_proj.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.410685505717993e-06,
|
|
"p50_s": 4.272034857422113e-06,
|
|
"p90_s": 4.822446499019861e-06,
|
|
"p95_s": 4.89124795421958e-06,
|
|
"p99_s": 4.946289118379355e-06,
|
|
"min_s": 3.999972250312567e-06,
|
|
"max_s": 4.960049409419298e-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": 9.290330732862154e-06,
|
|
"p50_s": 9.215029422193766e-06,
|
|
"p90_s": 9.74057475104928e-06,
|
|
"p95_s": 9.806267917156219e-06,
|
|
"p99_s": 9.858822450041771e-06,
|
|
"min_s": 8.78400169312954e-06,
|
|
"max_s": 9.871961083263159e-06
|
|
},
|
|
"linear.attn_out_proj.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.445330852021774e-06,
|
|
"p50_s": 3.4240074455738068e-06,
|
|
"p90_s": 3.5136006772518156e-06,
|
|
"p95_s": 3.5247998312115673e-06,
|
|
"p99_s": 3.533759154379368e-06,
|
|
"min_s": 3.375986125320196e-06,
|
|
"max_s": 3.535998985171318e-06
|
|
},
|
|
"linear.attn_out_proj.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0022702256489234665,
|
|
"p50_s": 0.0022497729514725506,
|
|
"p90_s": 0.0023021497880108656,
|
|
"p95_s": 0.0023086968925781545,
|
|
"p99_s": 0.0023139345762319863,
|
|
"min_s": 0.002245659998152405,
|
|
"max_s": 0.002315243997145444
|
|
},
|
|
"linear.attn_out_proj.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 9.96800372377038e-06,
|
|
"p50_s": 9.695999324321747e-06,
|
|
"p90_s": 1.0348809882998465e-05,
|
|
"p95_s": 1.0430411202833055e-05,
|
|
"p99_s": 1.0495692258700728e-05,
|
|
"min_s": 9.695999324321747e-06,
|
|
"max_s": 1.0512012522667646e-05
|
|
},
|
|
"linear.attn_out_proj.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002868635676956425,
|
|
"p50_s": 0.0028517399914562702,
|
|
"p90_s": 0.002900878433138132,
|
|
"p95_s": 0.002907020738348365,
|
|
"p99_s": 0.0029119345825165508,
|
|
"min_s": 0.0028410039958544075,
|
|
"max_s": 0.0029131630435585976
|
|
},
|
|
"linear.attn_out_proj.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.453308742493391e-06,
|
|
"p50_s": 6.511982064694166e-06,
|
|
"p90_s": 6.588769610971213e-06,
|
|
"p95_s": 6.5983680542558435e-06,
|
|
"p99_s": 6.606046808883548e-06,
|
|
"min_s": 6.239977665245533e-06,
|
|
"max_s": 6.607966497540474e-06
|
|
},
|
|
"linear.attn_out_proj.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.008090866671409458,
|
|
"p50_s": 0.007834755000658333,
|
|
"p90_s": 0.008461467805318534,
|
|
"p95_s": 0.00853980690590106,
|
|
"p99_s": 0.008602478186367078,
|
|
"min_s": 0.007819699007086456,
|
|
"max_s": 0.008618146006483585
|
|
},
|
|
"linear.attn_out_proj.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.562673322856426e-06,
|
|
"p50_s": 5.4879928939044476e-06,
|
|
"p90_s": 5.795189645141363e-06,
|
|
"p95_s": 5.833589239045977e-06,
|
|
"p99_s": 5.864308914169669e-06,
|
|
"min_s": 5.328038241714239e-06,
|
|
"max_s": 5.871988832950592e-06
|
|
},
|
|
"attn_out_proj": {
|
|
"count": 3,
|
|
"mean_s": 0.013308762339875102,
|
|
"p50_s": 0.013069211039692163,
|
|
"p90_s": 0.013700198195874692,
|
|
"p95_s": 0.013779071590397507,
|
|
"p99_s": 0.013842170306015759,
|
|
"min_s": 0.01299913099501282,
|
|
"max_s": 0.013857944984920323
|
|
},
|
|
"gate_msa": {
|
|
"count": 3,
|
|
"mean_s": 0.008683037323256334,
|
|
"p50_s": 0.008554771018680185,
|
|
"p90_s": 0.008906488574575632,
|
|
"p95_s": 0.008950453269062564,
|
|
"p99_s": 0.008985625024652109,
|
|
"min_s": 0.008499922987539321,
|
|
"max_s": 0.008994417963549495
|
|
},
|
|
"norm2": {
|
|
"count": 3,
|
|
"mean_s": 0.003726057999301702,
|
|
"p50_s": 0.003699098015204072,
|
|
"p90_s": 0.0038113028276711702,
|
|
"p95_s": 0.0038253284292295575,
|
|
"p99_s": 0.0038365489104762673,
|
|
"min_s": 0.0036397219519130886,
|
|
"max_s": 0.003839354030787945
|
|
},
|
|
"modulate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.010754756338428706,
|
|
"p50_s": 0.010686943016480654,
|
|
"p90_s": 0.010987435781862587,
|
|
"p95_s": 0.011024997377535328,
|
|
"p99_s": 0.011055046654073522,
|
|
"min_s": 0.010514767025597394,
|
|
"max_s": 0.01106255897320807
|
|
},
|
|
"linear.mlp_fc1.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.288003159066041e-06,
|
|
"p50_s": 4.031986463814974e-06,
|
|
"p90_s": 4.684797022491694e-06,
|
|
"p95_s": 4.766398342326284e-06,
|
|
"p99_s": 4.831679398193956e-06,
|
|
"min_s": 3.984023351222277e-06,
|
|
"max_s": 4.847999662160873e-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": 8.207998083283504e-06,
|
|
"p50_s": 7.983995601534843e-06,
|
|
"p90_s": 8.828798308968544e-06,
|
|
"p95_s": 8.934398647397758e-06,
|
|
"p99_s": 9.018878918141127e-06,
|
|
"min_s": 7.599999662488699e-06,
|
|
"max_s": 9.039998985826969e-06
|
|
},
|
|
"linear.mlp_fc1.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.503984771668911e-06,
|
|
"p50_s": 3.4879776649177074e-06,
|
|
"p90_s": 3.5775708965957167e-06,
|
|
"p95_s": 3.5887700505554675e-06,
|
|
"p99_s": 3.5977293737232683e-06,
|
|
"min_s": 3.4240074455738068e-06,
|
|
"max_s": 3.5999692045152187e-06
|
|
},
|
|
"linear.mlp_fc1.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.0017449889952937763,
|
|
"p50_s": 0.0017389570130035281,
|
|
"p90_s": 0.0017614722019061446,
|
|
"p95_s": 0.0017642866005189717,
|
|
"p99_s": 0.0017665381194092334,
|
|
"min_s": 0.0017289089737460017,
|
|
"max_s": 0.0017671009991317987
|
|
},
|
|
"linear.mlp_fc1.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.4506649070729813e-05,
|
|
"p50_s": 9.312003385275602e-06,
|
|
"p90_s": 2.1907163318246604e-05,
|
|
"p95_s": 2.3481558309867977e-05,
|
|
"p99_s": 2.4741074303165077e-05,
|
|
"min_s": 9.15199052542448e-06,
|
|
"max_s": 2.5055953301489353e-05
|
|
},
|
|
"linear.mlp_fc1.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.002230454992968589,
|
|
"p50_s": 0.002145772974472493,
|
|
"p90_s": 0.002364140178542584,
|
|
"p95_s": 0.0023914360790513456,
|
|
"p99_s": 0.0024132727994583547,
|
|
"min_s": 0.0021268600248731673,
|
|
"max_s": 0.002418731979560107
|
|
},
|
|
"linear.mlp_fc1.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.261320474247138e-06,
|
|
"p50_s": 6.27199187874794e-06,
|
|
"p90_s": 6.425566971302032e-06,
|
|
"p95_s": 6.444763857871294e-06,
|
|
"p99_s": 6.460121367126704e-06,
|
|
"min_s": 6.0480087995529175e-06,
|
|
"max_s": 6.4639607444405556e-06
|
|
},
|
|
"linear.mlp_fc1.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.03514027533431848,
|
|
"p50_s": 0.03508026502095163,
|
|
"p90_s": 0.03532921141013503,
|
|
"p95_s": 0.03536032970878296,
|
|
"p99_s": 0.03538522434770129,
|
|
"min_s": 0.03494911297457293,
|
|
"max_s": 0.03539144800743088
|
|
},
|
|
"linear.mlp_fc1.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.6320180495580035e-06,
|
|
"p50_s": 5.631998647004366e-06,
|
|
"p90_s": 5.657610017806292e-06,
|
|
"p95_s": 5.660811439156532e-06,
|
|
"p99_s": 5.663372576236725e-06,
|
|
"min_s": 5.600042641162872e-06,
|
|
"max_s": 5.664012860506773e-06
|
|
},
|
|
"mlp_fc1": {
|
|
"count": 3,
|
|
"mean_s": 0.039202652658180646,
|
|
"p50_s": 0.03935540997190401,
|
|
"p90_s": 0.039365841995459054,
|
|
"p95_s": 0.039367145998403436,
|
|
"p99_s": 0.039368189200758935,
|
|
"min_s": 0.03888409800129011,
|
|
"max_s": 0.03936845000134781
|
|
},
|
|
"mlp_swiglu": {
|
|
"count": 3,
|
|
"mean_s": 0.024196713328516733,
|
|
"p50_s": 0.024238202022388577,
|
|
"p90_s": 0.024260946782305837,
|
|
"p95_s": 0.024263789877295494,
|
|
"p99_s": 0.02426606435328722,
|
|
"min_s": 0.024085304990876466,
|
|
"max_s": 0.02426663297228515
|
|
},
|
|
"linear.mlp_fc2.flatten_contiguous": {
|
|
"count": 3,
|
|
"mean_s": 4.272015454868476e-06,
|
|
"p50_s": 4.384026397019625e-06,
|
|
"p90_s": 4.435202572494745e-06,
|
|
"p95_s": 4.441599594429136e-06,
|
|
"p99_s": 4.446717211976647e-06,
|
|
"min_s": 3.984023351222277e-06,
|
|
"max_s": 4.447996616363525e-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": 8.181338974585136e-06,
|
|
"p50_s": 8.447968866676092e-06,
|
|
"p90_s": 8.460821118205785e-06,
|
|
"p95_s": 8.462427649646997e-06,
|
|
"p99_s": 8.463712874799967e-06,
|
|
"min_s": 7.632013875991106e-06,
|
|
"max_s": 8.464034181088209e-06
|
|
},
|
|
"linear.mlp_fc2.bias_cast": {
|
|
"count": 3,
|
|
"mean_s": 3.4186717433234057e-06,
|
|
"p50_s": 3.4240074455738068e-06,
|
|
"p90_s": 3.4240074455738068e-06,
|
|
"p95_s": 3.4240074455738068e-06,
|
|
"p99_s": 3.4240074455738068e-06,
|
|
"min_s": 3.4080003388226032e-06,
|
|
"max_s": 3.4240074455738068e-06
|
|
},
|
|
"linear.mlp_fc2.activation_scale": {
|
|
"count": 3,
|
|
"mean_s": 0.004601721011567861,
|
|
"p50_s": 0.0045881689875386655,
|
|
"p90_s": 0.004628655395936221,
|
|
"p95_s": 0.004633716196985916,
|
|
"p99_s": 0.004637764837825671,
|
|
"min_s": 0.004578217049129307,
|
|
"max_s": 0.00463877699803561
|
|
},
|
|
"linear.mlp_fc2.scale_to_device": {
|
|
"count": 3,
|
|
"mean_s": 1.3728005190690359e-05,
|
|
"p50_s": 1.2575997970998287e-05,
|
|
"p90_s": 1.8028798513114454e-05,
|
|
"p95_s": 1.8710398580878974e-05,
|
|
"p99_s": 1.925567863509059e-05,
|
|
"min_s": 9.216018952429295e-06,
|
|
"max_s": 1.9391998648643494e-05
|
|
},
|
|
"linear.mlp_fc2.activation_quant_pack": {
|
|
"count": 3,
|
|
"mean_s": 0.005966662662103772,
|
|
"p50_s": 0.005981990019790828,
|
|
"p90_s": 0.0060327299637719985,
|
|
"p95_s": 0.006039072456769645,
|
|
"p99_s": 0.006044146451167762,
|
|
"min_s": 0.005872583016753197,
|
|
"max_s": 0.0060454149497672915
|
|
},
|
|
"linear.mlp_fc2.activation_quant_wrap": {
|
|
"count": 3,
|
|
"mean_s": 6.144012634952863e-06,
|
|
"p50_s": 6.176007445901632e-06,
|
|
"p90_s": 6.30401773378253e-06,
|
|
"p95_s": 6.320019019767642e-06,
|
|
"p99_s": 6.332820048555731e-06,
|
|
"min_s": 5.920010153204203e-06,
|
|
"max_s": 6.336020305752754e-06
|
|
},
|
|
"linear.mlp_fc2.gemm": {
|
|
"count": 3,
|
|
"mean_s": 0.05552383201817671,
|
|
"p50_s": 0.05537804099731147,
|
|
"p90_s": 0.055793770635500554,
|
|
"p95_s": 0.055845736840274184,
|
|
"p99_s": 0.055887309804093094,
|
|
"min_s": 0.05529575201217085,
|
|
"max_s": 0.05589770304504782
|
|
},
|
|
"linear.mlp_fc2.slice_reshape": {
|
|
"count": 3,
|
|
"mean_s": 5.546646813551585e-06,
|
|
"p50_s": 5.504000000655651e-06,
|
|
"p90_s": 5.619158037006855e-06,
|
|
"p95_s": 5.633552791550755e-06,
|
|
"p99_s": 5.645068595185876e-06,
|
|
"min_s": 5.4879928939044476e-06,
|
|
"max_s": 5.647947546094656e-06
|
|
},
|
|
"mlp_fc2": {
|
|
"count": 3,
|
|
"mean_s": 0.06617354867436613,
|
|
"p50_s": 0.06602008803747594,
|
|
"p90_s": 0.06640077200718224,
|
|
"p95_s": 0.06644835750339553,
|
|
"p99_s": 0.06648642590036616,
|
|
"min_s": 0.06600461498601362,
|
|
"max_s": 0.06649594299960881
|
|
},
|
|
"gate_mlp": {
|
|
"count": 3,
|
|
"mean_s": 0.014910061300421754,
|
|
"p50_s": 0.008941200969275087,
|
|
"p90_s": 0.023517421761061995,
|
|
"p95_s": 0.02533944936003536,
|
|
"p99_s": 0.02679707143921405,
|
|
"min_s": 0.008627505972981453,
|
|
"max_s": 0.027161476959008723
|
|
},
|
|
"block_total": {
|
|
"count": 3,
|
|
"mean_s": 0.5105496386337715,
|
|
"p50_s": 0.5060367499827407,
|
|
"p90_s": 0.5189121219678782,
|
|
"p95_s": 0.5205215434660204,
|
|
"p99_s": 0.5218090806645341,
|
|
"min_s": 0.5034812009544112,
|
|
"max_s": 0.5221309649641626
|
|
}
|
|
},
|
|
"module_forward": {
|
|
"count": 3,
|
|
"mean_s": 0.4635661483431856,
|
|
"p50_s": 0.4635968530201353,
|
|
"p90_s": 0.4642748417914845,
|
|
"p95_s": 0.46435959038790314,
|
|
"p99_s": 0.46442738926503807,
|
|
"min_s": 0.46265725302509964,
|
|
"max_s": 0.4644443389843218
|
|
},
|
|
"module_forward_checksum": -778077863936.0,
|
|
"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": 1357.5660000000003,
|
|
"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": 1349.8219999999997,
|
|
"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": "aten::linear",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 380.12699999999995,
|
|
"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::quantize_nvfp4",
|
|
"count": 2,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 333.52000000000044,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 5376], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 282.47900000000027,
|
|
"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": 191.07200000000012,
|
|
"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": 165.34399999999982,
|
|
"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": 150.27199999999993,
|
|
"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": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 131.98399999999992,
|
|
"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": 127.2159999999999,
|
|
"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": "aten::to",
|
|
"count": 12,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 112.27199999999971,
|
|
"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": 111.93600000000015,
|
|
"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::linear",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 110.38299999999981,
|
|
"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": "aten::copy_",
|
|
"count": 8,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 109.64799999999968,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::empty",
|
|
"count": 34,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 104.27200000000116,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 8009298432,
|
|
"input_shapes": "[[], [], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::scaled_mm_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 95.03900000000067,
|
|
"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::zeros",
|
|
"count": 4,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 88.20800000000008,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "aten::mean",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 71.7430000000004,
|
|
"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::to",
|
|
"count": 4,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 66.44799999999987,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[], [], [], [], [], []]"
|
|
},
|
|
{
|
|
"key": "comfy_kitchen::quantize_nvfp4",
|
|
"count": 1,
|
|
"device_type": "DeviceType.CPU",
|
|
"cpu_time_total_us": 62.83199999999988,
|
|
"device_time_total_us": 0.0,
|
|
"self_device_time_total_us": 0.0,
|
|
"self_device_memory_usage_bytes": 0,
|
|
"input_shapes": "[[37810, 7168], [], [], [], []]"
|
|
}
|
|
]
|
|
} |