opt-1.3b-random-t / quant_log.csv
iproskurina's picture
Add files using upload-large-folder tool
7b8152b verified
layer,module,loss,samples,damp,time
0,self_attn.k_proj,0.00409379,0.01000,0.998
0,self_attn.v_proj,0.00072114,0.01000,0.988
0,self_attn.q_proj,0.00414090,0.01000,0.992
0,self_attn.out_proj,0.00001473,0.01000,1.021
0,fc1,0.00906052,0.01000,1.003
0,fc2,0.00011689,0.01000,4.243
1,self_attn.k_proj,0.00241560,0.01000,1.022
1,self_attn.v_proj,0.00030304,0.01000,0.993
1,self_attn.q_proj,0.00263582,0.01000,0.988
1,self_attn.out_proj,0.00001348,0.01000,1.031
1,fc1,0.00911925,0.01000,1.005
1,fc2,0.00113884,0.01000,4.290
2,self_attn.k_proj,0.00382860,0.01000,0.983
2,self_attn.v_proj,0.00054063,0.01000,0.985
2,self_attn.q_proj,0.00378607,0.01000,0.988
2,self_attn.out_proj,0.00004634,0.01000,0.984
2,fc1,0.00967952,0.01000,0.996
2,fc2,0.00550816,0.01000,4.239
3,self_attn.k_proj,0.00358004,0.01000,0.985
3,self_attn.v_proj,0.00060755,0.01000,0.998
3,self_attn.q_proj,0.00388739,0.01000,0.986
3,self_attn.out_proj,0.00007405,0.01000,1.050
3,fc1,0.00773898,0.01000,0.991
3,fc2,0.02920380,0.01000,4.234
4,self_attn.k_proj,0.00413768,0.01000,0.986
4,self_attn.v_proj,0.00054294,0.01000,0.989
4,self_attn.q_proj,0.00418821,0.01000,0.983
4,self_attn.out_proj,0.00005169,0.01000,0.986
4,fc1,0.00736058,0.01000,1.007
4,fc2,0.39130572,0.01250,4.257
5,self_attn.k_proj,0.00486233,0.01000,1.011
5,self_attn.v_proj,0.00066676,0.01000,0.993
5,self_attn.q_proj,0.00465768,0.01000,0.992
5,self_attn.out_proj,0.00006351,0.01000,1.005
5,fc1,0.00789846,0.01000,1.005
5,fc2,0.00017804,0.01000,4.146
6,self_attn.k_proj,0.00761369,0.01000,0.989
6,self_attn.v_proj,0.00081846,0.01000,0.974
6,self_attn.q_proj,0.00610037,0.01000,0.981
6,self_attn.out_proj,0.00008329,0.01000,1.009
6,fc1,0.00947184,0.01000,0.977
6,fc2,0.00032562,0.01000,4.177
7,self_attn.k_proj,0.00821398,0.01000,0.989
7,self_attn.v_proj,0.00094582,0.01000,0.974
7,self_attn.q_proj,0.00643242,0.01000,0.995
7,self_attn.out_proj,0.00007262,0.01000,0.986
7,fc1,0.01092366,0.01000,1.010
7,fc2,0.00015775,0.01000,4.204
8,self_attn.k_proj,0.00864247,0.01000,0.986
8,self_attn.v_proj,0.00102178,0.01000,0.991
8,self_attn.q_proj,0.00637043,0.01000,0.974
8,self_attn.out_proj,0.00007326,0.01000,1.000
8,fc1,0.01218541,0.01000,1.032
8,fc2,0.00021664,0.01000,4.152
9,self_attn.k_proj,0.00874096,0.01000,0.996
9,self_attn.v_proj,0.00111076,0.01000,0.996
9,self_attn.q_proj,0.00673135,0.01000,0.996
9,self_attn.out_proj,0.00007552,0.01000,0.990
9,fc1,0.01299828,0.01000,1.013
9,fc2,0.00025837,0.01000,4.185
10,self_attn.k_proj,0.00844553,0.01000,0.972
10,self_attn.v_proj,0.00138582,0.01000,1.037
10,self_attn.q_proj,0.00662948,0.01000,1.033
10,self_attn.out_proj,0.00007813,0.01000,1.032
10,fc1,0.01408036,0.01000,1.002
10,fc2,0.00027341,0.01000,4.212
11,self_attn.k_proj,0.00810582,0.01000,1.027
11,self_attn.v_proj,0.00164685,0.01000,0.979
11,self_attn.q_proj,0.00655979,0.01000,1.001
11,self_attn.out_proj,0.00006866,0.01000,0.983
11,fc1,0.01486147,0.01000,1.003
11,fc2,0.00035751,0.01000,4.201
12,self_attn.k_proj,0.00892932,0.01000,0.974
12,self_attn.v_proj,0.00180916,0.01000,0.993
12,self_attn.q_proj,0.00685191,0.01000,0.987
12,self_attn.out_proj,0.00007998,0.01000,0.995
12,fc1,0.01568660,0.01000,0.986
12,fc2,0.00033367,0.01000,4.177
13,self_attn.k_proj,0.00930009,0.01000,0.995
13,self_attn.v_proj,0.00193084,0.01000,0.999
13,self_attn.q_proj,0.00691767,0.01000,0.989
13,self_attn.out_proj,0.00008400,0.01000,0.992
13,fc1,0.01652814,0.01000,1.009
13,fc2,0.00043007,0.01000,4.208
14,self_attn.k_proj,0.00859589,0.01000,0.986
14,self_attn.v_proj,0.00245260,0.01000,0.985
14,self_attn.q_proj,0.00656543,0.01000,0.978
14,self_attn.out_proj,0.00009421,0.01000,0.983
14,fc1,0.01786581,0.01000,1.007
14,fc2,0.00054112,0.01000,4.169
15,self_attn.k_proj,0.00779343,0.01000,0.995
15,self_attn.v_proj,0.00313876,0.01000,0.995
15,self_attn.q_proj,0.00688220,0.01000,1.002
15,self_attn.out_proj,0.00009604,0.01000,0.999
15,fc1,0.01899958,0.01000,0.984
15,fc2,0.00067443,0.01000,4.171
16,self_attn.k_proj,0.00774822,0.01000,0.970
16,self_attn.v_proj,0.00353068,0.01000,0.983
16,self_attn.q_proj,0.00607082,0.01000,0.979
16,self_attn.out_proj,0.00014269,0.01000,0.987
16,fc1,0.02137345,0.01000,1.042
16,fc2,0.00070578,0.01000,4.236
17,self_attn.k_proj,0.00760870,0.01000,1.032
17,self_attn.v_proj,0.00402974,0.01000,1.009
17,self_attn.q_proj,0.00628217,0.01000,1.044
17,self_attn.out_proj,0.00016927,0.01000,1.025
17,fc1,0.02438176,0.01000,1.000
17,fc2,0.00098657,0.01000,4.209
18,self_attn.k_proj,0.00715731,0.01000,0.977
18,self_attn.v_proj,0.00468908,0.01000,0.989
18,self_attn.q_proj,0.00624552,0.01000,0.978
18,self_attn.out_proj,0.00030485,0.01000,0.991
18,fc1,0.02636197,0.01000,0.999
18,fc2,0.00203174,0.01000,4.254
19,self_attn.k_proj,0.00734770,0.01000,0.990
19,self_attn.v_proj,0.00605974,0.01000,0.989
19,self_attn.q_proj,0.00633731,0.01000,0.990
19,self_attn.out_proj,0.00041913,0.01000,0.997
19,fc1,0.02817888,0.01000,0.998
19,fc2,0.00118158,0.01000,4.213
20,self_attn.k_proj,0.00741961,0.01000,0.991
20,self_attn.v_proj,0.00686072,0.01000,0.991
20,self_attn.q_proj,0.00603745,0.01000,0.996
20,self_attn.out_proj,0.00052878,0.01000,0.992
20,fc1,0.02827213,0.01000,0.996
20,fc2,0.00133380,0.01000,4.248
21,self_attn.k_proj,0.00697015,0.01000,0.991
21,self_attn.v_proj,0.00690806,0.01000,0.995
21,self_attn.q_proj,0.00655608,0.01000,1.053
21,self_attn.out_proj,0.00054652,0.01000,0.991
21,fc1,0.02712614,0.01000,1.007
21,fc2,0.00177069,0.01000,4.216
22,self_attn.k_proj,0.00651944,0.01000,0.982
22,self_attn.v_proj,0.00744933,0.01000,0.990
22,self_attn.q_proj,0.00703869,0.01000,0.977
22,self_attn.out_proj,0.00068570,0.01000,0.989
22,fc1,0.02461607,0.01000,1.004
22,fc2,0.00199070,0.01000,4.303
23,self_attn.k_proj,0.00893416,0.01000,0.989
23,self_attn.v_proj,0.00480653,0.01000,0.994
23,self_attn.q_proj,0.01262821,0.01000,0.995
23,self_attn.out_proj,0.00080622,0.01000,1.000
23,fc1,0.02185187,0.01000,0.985
23,fc2,0.00234254,0.01000,4.236