From 6ca30d6bf4965ddc410dbee80ef04d8037298722 Mon Sep 17 00:00:00 2001 From: zhangyi Date: Thu, 13 Aug 2020 16:50:32 +0800 Subject: [PATCH] Modify mindinsight tutotials for tensor and scalar. --- .../mindinsight/images/scalar_select.png | Bin 8262 -> 12307 bytes ...nsight_image_histogram_scalar_tensor.ipynb | 15 +++++++-------- 2 files changed, 7 insertions(+), 8 deletions(-) diff --git a/tutorials/notebook/mindinsight/images/scalar_select.png b/tutorials/notebook/mindinsight/images/scalar_select.png index a5a75e646fd236b4350b6f5dc57f4fa08b2096c6..07a7850969f2193cec615012384030f3af07209c 100644 GIT binary patch literal 12307 zcmdsdXH=8h)-LL1vkwT0(yL`#=(6c8QBhF=Q6TgpAWaAX1B4QS4N+^-m4;ov~U^Dvz_+2{ZA zy=e{Q;NWZf@ja9z$0yFgq4eR_wJYX94y%(qZ$Ud*7xtttzC6SIdgIN|CjTW|)D7#C z(RsG2C7vfT+CHSAR@8C>pi1?(+PaT6uT^I`$dnBK-Q7^KJaz4aN~MLxVpiOoe_Lw8 z$nzNfMli~KxaoJ8McRc@cQ5-DFVW}^RtFH-cmN3hg}z2y z2d-UaclPHe#*egx{C(__zH##B;p6lT>cKu;BDV`8+pIn|b@=DiO9iVFjT*_i%$rRO z5QuK07Yvi!JA0%sTddGX6ENg}`LF@KiV39s3wY>fdy#%I#AbgTWul4%C1)m9IH@*} z=wwo(l0O!xc<=?@wb4%%`t?3#xtfN_dI<76^eC;BF(q8FG6^%X{m%VwNmX&uKtT3n zjhv%%wPD5J^H#6G(!YM+e4KuY`NvN&91oxV9stLOf1N#nz|@4oR=NvnH`RXK<}h%d zc=8mEx!vH>wRx_$GUw8GkA#<{V(jARUaj-)TJz{ck5P<_*`2RsofGq>Qxlbneh)hcW=huz9Jn(+9yesw zGmx#ctt^q_5=={z1+qf?AeTb!N}@J5Ge`%ja=)Jtffn;4e8RS$q&B-dNYHY(cnrs9 zrWoNuu@Ir!9hi-G-N7+ZAAGTgk+QAlb6g21WU069%)H+xike%BP?mEHvqDh%MRq4< zTPbARUaLzZ3@1qWWZ*t49xpjfzWbk^Sr#K`!F5s7?!(OG83Hh&Vu_qWaa^Z^<+UHs z(x?n_5Hm()rcc;Wo~hf~Q+a_Y|DrISH9O_;@|e{0mUzS1zkr5d2r)!)t)9+$Wzg4< zjjo-fjDTU%2wyFT^#sCen&v%SHqsmlQHRZ*$Gw%gFuV18mN@bM-0isfS+&xozMoIy zZ;66-f+^O0@nCCCM3qC~&-~h1Lw%LDrT(+#I@9jA47}vMAI8fUtQ1wRyd2^*R=#P! z@V$@H#jjvsOBpqBsRrBe@6b>DlAUX*CJT%&|7UYS4j6iM1oPR8c)9v&Ls$0)#t6u7 zvJ(!v6*!Si+j0@{=gY&j5^85&RoF`*8-W!)O?g^~AaY+SRCSs>yfpLrr;Qa)-;*Ae zcT{at*8jb9Zs7sUPPvO)kCnZiFlJ7-Ys$5L&HcbMTJ=QABxg?rPM@0gl;Ke5v{*aK z)I|rf{WZpar;gdkD z#icY{z-2Cm;pSHyPIhcjYfqYF2nQ4jS_*^X%ez;x7L6B>qFeLkQjN_eD(J#kJ#f%( zJ;=p@-z68TBVVPWE{N@RMk^%hL|-9K5OH;_OXfw^mfXOOp`AaYuu+{~)I5^NPK%E8 z(Q%AM1hRQNuDoo$M{;=s3P0Ey30+3j71o$+X!$uC83Dz8qJY`by{M32INbywDs8SYueGiyA;fscO`>`6nzYgv zOA*K{@iXE}&>2p@&rLJMb+gHJWXwEmIU^ajU7P(OzIBB9-!sJhvGp)~RE| z9XcdRsziVu*t=FR+hiJjz7X(1<-O0NHd02on5ABoo=@Y)2A6Otxj`Nvd-_uZ5OyD( zt?kT)-cV;7@=+d$Hd`R;7W2c~FhvKE{Yt`9I`44Xinotp>E1m9b&Z4<^ED_x5^l;9 z3{{;Se~&7AG>)t3o%m1p?e`4!^*@)z2jUbG-P+zC!r?M^Frt(qjauwp%9fzs3$~Z+ zY}jo(cHTIy0gkFLBxef&M=*MbyFn7z+x-LQcc0YN5?!QqiiHu~MbW`0h4Hb`hOWrD zK>UlrU19ue3>M`-eYw~DGqaTf`+@>u=*Z(>)=@6X^T5 zs<*v=?MaWH(OoBYW-|Q7nLiw-nbUkLZL5DOCKO3zFIlX~R&jTZIYMZ+eyNu^lKa+X z3-J(F?;2mF)4&Z2U91+3po9I^q2pCGwOw6uRt!*7oF0i!V0H^{1=FRBc=>&6O=NFo zy1w2XaW!%&xa~>cE5Vd-Eh(AO^q`$A%7&xei3h0Rv%RyL~bdY?{WEN`efERVH%kiV8PhFlBaEy?7-H)WUbHpO+ zHxSUy-?#)akuvn0=o$^3$0mQvQVt!D8F>@{p;?b5;|_P_ecG!3Cdk+xTu*)Wm1^v? zn3H$AFMHe&l$mvJ7gaLTu+~ZnE)em@<*9JTdp(>;Y8_GqhMROE+)S!ds5)&VfDL}H zrr2T5-6-Brh|kWPF|h=3`|RGZv z(6oDK@@2GvC7J4MJDEs`35N2Ws5qb~M2xqu{q=J(K2|;z5g{|;vaK^zF?jJWAbVnG zUIZR=Ek8gh2ZzX!=SUZPqb$>?!NO5`cq$qDw)wnbT+6|?;rFUd?h$MNaaT;l ze?;}hI8HF<;FpuTzkL!BC#fciC~w)Bw@TXV3htPp+6sf7BO!XIf6{HX9BP$y6g{j& z!Uuwk?MyLd5kJ6GY0m1UbP=-x3Gdl^Ki|!9gJ+sNa_`yngoBX=+ zSlO8NuV04)fBd_d%MfOtL1LrKbS34n{9)N-qs|Yy zRerua!5ql^0W3ON6XY+lJ#kObKczRXw4XY3!@$dl_4#V@npG7%xU}fz**J3{W zeqzXK#A!QHo~O_xp0&)}#VbTCx~RC%dKp|6vMC(ixc?2RyuH+DQameHCWn-k16!uvl&Wv_Q%m>#iydfbk#;=E)om}b6CN$BUNWN% z4or!FW&D5a9@&ZIa(!?XSdJ6mJ?$KlEgoA4)&mgTAO6UrA2Lb}_S5t*}=v)=!3-0JO^WNV_vj z#0+uh9bILIa7Gr~Fv*o%q9>M-k?Xs8%tk>cY_al#m#0igw{a4j7e#y7xBPl{>g1i5 zz-*NZfKVzMO3ufmhRx)`4FmFf>ULom$PTnP6)-||J{P!7#e@?kdj|)PZ@H19_kAh| z%A-UZB;%rO3MdTi{guVnvAZ1FxKFffXl!_&Lw+K$6fC%20t-dn4|~_X{%n6!%9*u- zu^HahEkFs)9QmQEPadYt8&!zw$_brvXQL0z9oKJNQ|s+BwbU&57|t!35^%#%=(vD> zj6jrPhiUho70S=ARNS|z*?fOgWOZ5ljP_bwGVk1JPi|xJ_rYM#`V~|@!?0t~FXifR zk|9s6MMAcj(Bt_hH_QZls^onoBkpJ&&}cCCY<_8^ijKo%t&-G4`u>u=Hl>WZh^hdYB??dg<*ZA-gZEgqsHk!ypwj*f2Tu*-g-lSBCgz%a=Z)HTQivh(@S{;2{r~x0SHo z75ZnS<=*7&(_Y&cX#o?X1BQ8BODnFSbsfA&TWQW|_Dmqwo?atPoyh8l8n=v_*Es&u z8b#)cr>N>RpI4OcUN>(kQq!*FsFc50=EyH6t}YN~e_~mvA~x2Gj4jSyD%-QV$9;iu z+sSn)H-i&p;Fy&L>T+ETG<+d7D`DuK>2I0|-Tda)N^4Jn**WbEO!QYp5fbBc6c5(& z6)4ZTT4L=XE%p;RvQatXZ{#}_y}UA;an^5iJp}EG54^n{g(V|clarOFdF~iKF;&i3 zUM2| zrah0Js;SPnzw-6b)uD=uFNznnBaN>A75agk^t6vRf|x%)6)m#Kb1DkybR`6tON~qxT)6S*{NbsN?p*qMO0cTmH4<=kCuzW4-nRNPk zyh&$TRz&8xl&&BMd@YQOOS@nkhx8d0TeWX_RjwDhS%!B&C3#3$9TNcvKYNe%5SwM@ zuP)XdI-GxRo|s9^u3Kqdk2O1Q^mSMu-;x9rcE~U+L<(g%jGsK_JtZODP%Erq(VyJi z3W3`XBpt?dZhKfw92eKFeXvzzstvlx$7E#6`tdNR=Hpa|T4R2d@3ACUf(%5))^GI_ z|IJnJTqlQZ|Di4c{lW5b=AlBejb%z?Mr4Qs_Acxl-oA(#mZvF1`7G(oLJ)rKJDMjz zyg;*aR`yUSU$?Tq*Sm{prm5CxDts89&rdH zENTU`K2Hsuc||GQBI7yP9GOYe=4}fwZ6HpPeh+Wsuv+iqR}K2Z?TCrYaW8LMp)8r0 zI~6XTO(U&uar*Md2)Fw$gCShpv&E-ts`m+Gq%^U9o7k2D^hFo}r=Ppj zw-uo8se=|_hnEai02JfufAfiJd^T~e#`?MkW#;kJ$4e!tooMo^mt1$SSYl0OI?A8_ zYSEI0WJbVeslcSctfH|d6!VG1Ot7TgV%UeK0H4KX*He*=Kj=;l4qsJvFf>mJRP_0p zVp!(Ag-%LRkd98W9i?U78j{iwzpGMt*UwX?X#rSfb#DS4I8~g(DHd5=CtEXbyl4ZM zCr9m>zI$zgoh|z(MzZwbwn!*qcP;4t7KB7uTl;_vdC0_X&Es2uC5i{N(C#>Fl{?n3 zqJ!LZ zR~=-;8mcuQj?r#1oyjG=D5(3;rh06}UJ_3?!G>QcRb3%DE7Ye6HuEnE(Q%=DjBkCc zwj$_mO;@%sZ?9nf%-8qM9wV13)*Y`cK7HbBW#t`hs~BFkd0drIWgI>n8{N?z&aH1) z0a^@LBtAtWORom__Af7Shlggpeh3OpvV>5NIfe3WH@V@PTG!1@20a|})f*S-$%=6c zikbKsyYT<0)SRtSpT(F4^nsIkar;9LOb&5xPdq(0Ck&+ABp+L^5o#=7@t0tK$*SMltXfO)NNSityZ8qp)=zy~8X)ZWt zqn@|3sB{L>Z*zc8ia%*H;n2i!3N%B#*qCs-Cb#SZx-%&DW%w=1_|@DC8C;b&6~wB_`bce_y>yGZa-C7U$>9$2=RICv zA^Sqqcr~_H{pQWIv2bd|V{vh+l;-i}+I>YEWSF{C2z^feamGJRSDC(B^R%gGmRL&| zDHcw6xNN4JQ6TWgj+c{N>TU4zV+Eu&kI^33M7x~@O6MOo8>yJda$Fq~4Da8`ph>p` zE8D{6GPSH5$Gad+*`O&vGt^zkJU-KT(8bn%y;uFMgwyES^F996^07>CaD4RW%G%`95R6Cn1}U4WBUSu zKq0j7R)_!T)-jK@rLfFVVt|8vx((P8kf5(u(JEzVDI{fSIfc_O#`-<1qD1hlyI>3N zjU~(WMGJn_OIn)agN6Up;r9Ssj}S4N>d-=Y3;;?kA5Y89rXy7--FaToof%8x~QCEVo} z8(Ine2S~Y0R-=0JaUwm%nDT`ULRx}z5P7QV@z_?lAra{`1Vo0LDkBR15o#s|v;tYMVvFh&j zee>TDt3=vTBvRs}C17)xvVbOH%flF(YLVnTQhfY%75BDyCB`#&*CoS4llk{7tRUP7L<$4 z(s+`%Sr@k|DrHA&!^7dE`fl1fJ84$J_SO)Ga6O`FhGMq5!U1((x-;dzzxDx*G?RDI z}cSvhsjz${E$fWb0o-jf<~dVX6w;UToTKVd%)On|7mvrD~t3$c$mx5 zI=gGp$#F=0A?rYTrc}E0729cv_hV|c6g`j_=C)#XaBU^nDebB9L4WLyEd(Q z#$GGMI{ZMe?)%>({k-r5o~@QxHXg3r(b-*%-Wf+@huK?jhqUaAcQSwfgz5LUMPTnv zA*Gci%Xpk8XJS$B0glR#xAW<5l6mQUJ*uh|u}Ox7WJ6BLGSvOSCacCro^Q((v`5fR z3B@ns(&(0Lajo@@I~XKZV-qwtR^FuOqpRll(w?$XQrH%(*KwWT)1WcM3c_59`RQ3< z{@KiV@=R|84A&)W;2pdnM-eV@?5;K6>k*D-9KGI%^7WxEY{HxKp_!hW9n{l|HWwOtsj&h2gd<23Cy$K+zmN6 z9=}$0YbAVskmEtFoDp7(OjvUtYyv#-_?F5I1WL3t98?D}v$=9*6r|2DS3uf6p>`pw zvowe&%Mi%llFoGNOxHSFFdxe6emtbKmGIjm#DeA2 zt!h&(?OjI$;xj1-{rsrzannsx5%xnfkg_Phb;c6Sh_|;KU#)}NHTc}9Z|YWcU^i?% z_huqSm)NXafw^8oKmar_132}DnkqVckTRp}*y!dK*m*i%$q_^M?>uO8ivnydg=c-x8U4ZrjDcy(@3PYcO6D5wDn@|n;a@Lc%pUcBTT!bDcm^*NEZifQog zl|;hRz5y|fns!c(7sSMp7HM1HHhqflOgKDeLv&}Z4zklv|4}3I!5q7x0%B|pL zvmF!MXzxj}s_yJ0qQ&~#t99m2QhlJ`OK@g2ZgXMVFG6HeJ799Ru6Q8#GXS9m+mqmd zpv?o?j#?30LeO>*4=LBrpfrZCYvCZ2P1x;Tv(3@^hm3~Qr5h8sBZgn5VpNj{1*hbW zi9HsYT2HCIW=|LQ*F$&9h@hdI80y}V`M8%9Wt4*RB!*6l0RWEB~w+EvSo(mGaGM#?=L8qXx(Q36)#43)}G%vZh` z>Di5@s8a1hbOon9Y$dv38!n%_HDK2|krInu&q~U5<+|lcn7WYpn2`ogwp+M0O`g}U z>6ScXGKH)$GEPF~-z^m>1^ww$bQ`tsp4Mp~6+JF3Pqm(`T%1eQ^RgJSPy=c3O#ug)egJvFb0uF*NMXtL z{%^{!ctUP(VX5d*IFVsx8H6-pxVuE44a^9fa$#{D$&44d&q71+-7DnxY5Jvm-S%oQ zAL>>bGl=89COFnn7~ot$$k_Lu{QE}=htb&TE=K9vR7nvb#)Xi0GIOriy-i* z+%_|IYgA{Rykcc0p<7;?<$SC7c)NQ3Mioq=K8kX+A54v*r(v21_`Ky%TLTX(9J^}4 zOKDLuW2=XX+qQ8+<{-nSu^O#m%IJsR#-5U_ru7)GO-#31)r{V(Bji_G6^u8k~*-BEKYat-mw?J_>-V zI^J0YK@M)d1^rFElQcb2TDMtXf=`C1;lum_9h~f}U|prpVvY@j+qI7(X;$VgO{Kcf!gE+v=qGT#pH6=Y1smAtm0O=WfwQ3*FR%fGzV^3tX>ei{@K0KwZD{RRMcMC`FVV5 z-c9KhAupiHeFtrcp)J(~cDO_;^SN{H3E^^yE0W|&w+E}wk*|`hgP2LW&*c@ub+U_7 z*T2+sKoEU0A_JXGv!FR$n&lTp)bpBe6y6+154tA6=3kT=25#=hF*JIN{*{x-}`|_CAS9_~WUb$=NiL$^3 zPPPtoYh5?U%7l(G+6Ti1i?m3Rv%L?vMd)8 zYHL1qxW;jL)XsH$ym}_i!$y*lGa^QwUA5}ifplqGjSo5O55@1GYX$!djX6?U(ON%9 z4LmCPk#qI(UUh!g)H+woKwc$h$YPjQta!)*k2&8cpgDfpBb3ovm`pW)!q&m0Z&-Od zfa}IVy^=;xL+{sLu`66Zz)~#Qb&HTjK=q#SG=6oUCw%iHp077OVvBB zb_SisO>W&+&Ht`aX0+<7VG>nQxYSh-aa)BDSwS*1Bhb5HZ2ovO*jM_dz32r;ew}il zf)s&j$XJQmfn-auD{|g#oZxlb(5lRi8)y4p_(Hl;TUU|fazZhWV;tk4a^&i+l0{2q zg^rOR7}BrWfIlt!XCjsi?|62^?jExKqEC5{)LE#H+}Kos=iBg6EJuSzTLW(_s!sQ< z#KSpvn5F-9`vGH<)j%Fb9$;}l4jet87j|>CpJAn1H4+!Rj|Ptta>nXmjzVxYlzaO9 zBuCtn=n6L030WHV3oUB?nh0xaFLcAz4AbPm+CFTtk~pTwR?wrxRPV%rZBnSqL_DkLlI2mgP@AFo9CDrczV(FKuxJX= zl3SI0eezYP%ksP`j{J<@QYfzRkq_E(SbrC9UvVcUdOkUpOFK#u-boNx2-88^^_!L+ zZl>{|;8YjN?CFyY2P9Leg>Yso2VJW_6$a;U`R+(Y4Re*H`-gL>Ubw zqx09^Dfq@Wtb$u0BZgw{g9mDw70|_cN{%6c{dY@LJL1M071qX-)qFCwl!1VV59Axe zVaINmIc*^kD@%212&w)To+|`?&roqL&u2Eh>F9ws)Z?;ow@y2ZrYLZM>(igM&ev6# z-?e_&AM!TNl~H8J()m#(Y^I}G_W*%4WaN4JTuW*SzWCgael8W2#mfRI>vhY#yg9(Q;2!PI-ZIYys?{#bou5Rf!y9a*N!4hNadW4KrV)NE0SEpE-;*|EXv}RQ zOJBFy#U1`|ZISZncbEY$f5gY`&5ib#5y*A*R+AUeO84R<$;nE3lm~1M!+DMYf||sO z6977jnd)ld1Qk#upbBar^aim%O0J9|s9Zn_n1nM)$^kRaS4248eI?8O;70D}{RJ$! zM*HSV`8kF+x*vm~6Lr(bnZX}y)e*y_oZ6vb9>Agd8~{r#HhIEv*7XOsbn&JOEHG>- zvDYsA%21y=2zI`ZJ&GU6ETIqpdv~5$PO+1x{jAXvaWzL+ylQ86Sng}QBw2{#eh|%W zPJ%5eMj+{bE%UJEWE_)tS#<%Ksq`zg#DRaxCM?3CSE|M{ubc!n$5@nSQsd2CDA1A7AfnIh%vh& zf@&Fv&DMjog%;fTHhh@FH;+$3^LRNUWE z@zTO(CR$nr#(G*m7`)3he*q0{a$k)%<75Xp10S|cx?sy#*Y$Zad!8@#{kH<(P_5RcU1M3g-Bme*a literal 8262 zcmdscc{H2d_iy@k@M?>;+G?w)w?$F3HP5Ak4yGEbrnYLRA!rZ;X{)8JqLVSk(wL>F zL=Y*8DiSl1sVE^tkPt-TM&I>a>wDL_cin&Q`u&kS=Q+42beUhvsDR*Ith zhuf*geQ~V8M#YzHLnA|la=?pHIK&V`ta896aN)}8qd67hVcW2eL89BlXJucA(K?c6 zeh9Myt8a~vH;Rutm04IJVL@({S6exFQypXH+zQO4A_z#~u?shzpG&DJcrLI40B}iI zF;(oQWXr}Mf?2Q0d zaB>WzXWz-e#a$mRV#zFGG^Tf4Nj zBY8S!`EE2LsZK956&`~ksgISdwm}VZy zAMJ;w;bGC~;a9ReyP24c*d^|-b8o}SOZ%MS2yE0tE&7@GB)WqQ3vvyt%+VlBsGV zaUx}nx;BZKOj2ouWa>@sp07R1rH*B~UU2oSimt6`KP6PbY1qnKp|<$8NKbT&e)Wv` z(H%+}@5i2aocU2^WFJwBoBe$V*jqa956o)%b3~iAOy2aOf;1Wa)VhX!HKn$+LbLVNeUbh93*)0( zfg+E>-F^7q+@xH7OWWQMWBie#|Mr^bc&b*Ju9bZh6q*cHs}HHTdqQ=wec7qrH+J(A z1RDunpWVE^tSG1K@w#n!|E+-wSg%JpQ(?vC^pZ2$d4DW*^YVCWgj+}Gr)u2u-;Yxe z0sVWW`j#d;H1#wJL&(Gur+lvuxGrV2ky$UpxwYnp#sB1JzH!#YpL{R$=#1F5AZJCc zu4RK12qhcC((3)67Mewe)Z5-1^dKvyhE{u!ZtUH(BH!FUuC5U*v_0b$<>esNM7lO3af9*%#*T83gf>|qFLV4magh( z1OU#vA6EM}{yb^?hkoGCz6b33dpr0Fc)s~J5M28HKcJ)BzdLC1tchPv9G!2c}v^XnkDiJwS*OQHpSv>rx2wmd$ut@$T5mukYg{RTv)hs~w`gx&jB z_t6AeujM3dh!k0$&*m?hP`#{LFR%0kx>!O>P zD0eh$J>dhro4Y}J%Rv`}wksADud7R-OD2|$iZ8a!VL*TT`+ z)U_m>Gvw4YdURi*nhC_o`+CJz*S){6ve`srT1wB3=1`+B3WIcG4J(c}*I`hO&WCpP zUouV1Uyr$T=M!A3L^E{4xb!K1N3BL7QtxUgeI`K}el*k@)4p(2arl5b-tKNp^#imt zdfJz44@Sdm_iKfWl!fY4HZ7cxJJGT4O3e!Tio(wXnHT5vdeyG74U?A9g6;dn#%?ia z!O_klTwNA&s&6LIs^UkgowtYSYp}9(f+8hNv?JiMR}AVnYFyXL^!|Y?eV5*{Icu)e z)@1dq(D{O7_9B$u*Q4^~VG#_e&#_v5Td*ijUmly;W;Hzx!Zo=yxe?VaO|5(2KRFqQp6%rEDT^Ce znFjFWq?a;(^9Yc{;zBwcNL+~aJ7thQdLUzdj--=XGiubam(>`S{tLa=#F&{8vNrF5 zJ)s}6S%>1I_*3rcsYkv$CnpE6S35A%A3E{eEQjBOG+bx2>I-p-qqvB+A*&q4?-MgV zS;k#S8r&p!H_drbH{8{?nDV^ruFTb@w0!(21H-kEctPNe|AizoJDY( zhvC@C#+gNs_dcv8hflOo>u%gquH`c4IGa7*O`KHv$kPnh~efKuSc3oQhrAZei7Q`ndWQD+iS=hl4KtjKf3c(eb*5#`w(x5 z$6295Ich-Q_!<{=*}}pVh+Vs!D?C>f?%v+moFuPiPLOAgbgzLF8*S>=4r=Kc_?voE zl#;)C-$?W&71M_y`Py}t-d8oq-}O4fg*;|*#G+LxZ7H+Iq=vt&Csg)C^<8u_kefs+ zgl9)kp-CY%U_=D(iq|A>`#9ch^E?Qs{7BPnh!BVRmE$2@jXc?Fs_EHF9Jh#lUu zHm?*=mLWc{=mpgaOxhP};H@9*SH&r5SbB4caGRFXMFHhRKp}L~rx@=c!_N1o(!E@- zoGmotvy#VntHg~2(9w)1F&I@{?m1U5C1S%Jx{|grGnu^7NTHOWjLCmWuj%1uw!sjph{cH$f7_MVPkI+AXI*_rnasW|noR{A%kvAMTJ&>DzP44i>hCzZ^T(^fh@t5B0>m2P-8<>7G zCw|eeyln|5Br}8MhZTi2EybxCu*~gqgX3~ySQflRDxX|&`&=mfIztHhgIFHj60JeT ze?EV6ZX#Y>W8E@(yvOL)C^lOQij6dU8y{)P*$8{Nj>)t7(yb+dC!oCC7QMaH-aLmz zb7cv;Tu{$2zqDe?7WFE$sV{Pd*`VZAc^gg=&46=-*^Jp5#L(kLTI5D$#(x2k>i26_ zg()(EMO*LPD~+4etJW(%QESgipIASX;bL_lo5i4T+kVrJz zm}o7b0nG=CiF3bII08==qi~rH7R6WNMl#!*w&ptmD(!PIPmb2VJdn=*vp41Py2A(y z?fZT@H2Pc31zb*yQSdYQ_JtFxC0e2~bQx46+_LK3J&;c^LSkio-uuR8Gfo0AaYP0) z&)eQRwu4Ajz>nd6`)g|BwNx6&{J2jPp@5VC-V4btyQX_jEAKSk+1gko0xe@a=73=y zFFe(}MwQm;EZ%st(T6RYco0FRGE?POIV_=U{8>4SJ!@r_TrY=yT?^fg9Nhs|#b_qs zHTPAA_74ecE{bvUu?ojDTBR`Vl;>kzTjyNsa_Cy!EY1)x4AHTAGFL$-^^Q(%sgqdp zz?zFch=a`k{?#bWl*zRE+4d)W208b&=)%2wY_0q{=D6g+PN(Sc-D5EfN7CVCMRdZp zO8}XLtV#Bmw_8EKG9@MC`VRc~$!DrWtp46Y7j#+kf zs=AXL#ow#F`!Md=Zcim+hMoO6w?cKi4!K8A4>{Wy8&4V1kEeodLBt6V-o~>{7r6KH zvmm2e{^z&msO-6w0yAP9YztGIZ5p~Q|EHwOb8n~;rjAhnP{|t@*BxaCsz)s zsQ-izJbirMO7kl*@?k;$dCozccc&z1MjCX=ca_3QMo(@RcUXUqpAG;gR#w90fR@NYy*cNhNx1yTA8dwJY~Z!$ zYl(#I0{vA$53stvq?*Vb9I!^NNC5qhs$ksWySuQggP>}<*Y5feQn?7N{*^eBE~aOe zs|+3>h+^T@3@k`|KpnZ@rLE;E^pyumF5OrN^h0Y)kDL}FeR}Ijjj%?kg2ly;5teou zx`d8=9A3%g?szo0t#*7>GPN*pE@#M|vIa(B#?^jY92!y2af~wyd6Wq0e?%;>pz7zr z7JwhIa?mYIzX4`M3%g2IAcrK!5<1YYkRKxEyOsZd&Dsz#CEDA_Lrr|PdP#1!3v{V% zEU2@szWP%O40(hV?-p(0>;~z9@i(qY(5nzGlZiV4*85qrq*zvYzg^3?0e0y!INk6< zz@W8rlXz#pC`~H_y8cUzy-w@-zH8ioacd)465qdNd2@^P$#|A!b=`^*o)ITQ$aiR1 z>naPKDNnOs`rfT{yG8~VyY%&x1|{T_{-SICYVMT#$UbO;jP;zA`?AEHuBOimiH`Nc z%G&qtj=#;oFK`3T{8>}k2Q)Q4m1?J&2O z#+%JT8y!NQ!oK{I z6VabBU+%aX;dzcCQUCX%@IIO{Zr}m*9rHV^z9%lUVZNGUzqIT$3$>kV^0N$*px?w) z8XifR%`@7mE=<$C|Cpd1GWnxyGLF@KiRPGyX!1>~U$*xtYs#0=#FknJ4R4tA#r3#O zE4oG9IKHlVZ-ja_V9en&r2FYb?e0wbOmPbG=IMKxJ2!cducXFda`RK8khi&OnmA6h zwwd{)PWpq`Flelkm2L_MyVuB4ha>Gl`%MJqbeOj|v<29P&2>sv%4nwx`Cq~=yrTM6 z2`^j4tZdkYN&%jjW+sBfn2nCj=uDZxT+B~nV4Ruv6|ua=S>Nhvc@ z@$Cs)AIH2`KXoJ+*G`K0pio9!ATFal`26R}(}#a~yymQTd|x=K?dQUie?pur5Mhr` zdY(sgG)hD3qxTB+y2oW!TOI9IX-IM%p(IEf90fdiuial1@2)|!=rQ@+c+)Qv(fgFP z>hry4hrG@2pAq%GQj%ZpM}6DZN0v82u9^o?3q;mnenYe(f5!{eDS1&n8nHB=0c#Iy zvan-Uha^}m7oOpOQ(utHeS?YLy!V;v!sB&mI7C>B%}!)>U_DyuQJbGUVrt1fxk(vX zepZf70hip%-U$E)(RXfv3(b2s-SE}x`*RCtf-BMtOj0A$|9WsO#n63VKBZx^Bf?R! zCesVHlO~l<54=`OiF4-E*%b359o++t6CElT5HU;>+YiN5!x$65purm_g#z|{;_m@F zCp$)1rN1;hC`lq~?X}i*U%^|=lxR~kZH5dE*x@}Al-Y~RYW0G1V9ip^*#4GLXc8<_ zO5KRFh%5UTaVy^tyunv><+XM@?<#1IaCoEy3}kew+$=V^0Ddo#&@t5S@oYWFC0aat zhCol&mF?-qhtiLAiw2cXc|F-Y=b_);S=ewsr)Jt-LQS zGrwLA9IS9;^uzO%vf8wYaK#kWEhI2I|_BbnC&Shr!R*pG~eUA)@v$59(CN#-XlsFJ!(q@=g);=o9!jh->X%AIvP5%r?u{>Kmo zGxk4uUQ8xEzhV|g<}bw6)-yLfHd@DD=q)Z6)B|&!@5xALc#7#IB`g*VhB0T<$9uzk zv}!41vGCrou|3w3fEt*Yw|0-EW-GGiRaiwu|#$duvdY9U{k_Xv;JR$171o>omyL81+JK={@qVDgfYPBmGvl6JbZ<^MZ~a zWqQ`NHeH>~>Cal?K_kcC-Ix})D@rRa>DtYHlw}Sl+zhWn!=cHZjspt65cjEp_Tw!m z5sFC-o2x`wg81S0j;>0tumjU%23W)ZbuiHuf-XIp=@h?8M`~MMH=`!6#juvl2g}J? zQ%J{S0l(NVA7cX?um5Ql7j9*GZNOi>u&kdp*Z&f$htwl)puF6}5e(G|VTtb6d3z@W zL7ywy^bnf-K%}a7cVCHypH(SV0Unn(APUKaxZJGEnFz4QQ07kIq#cVWq?tJh*BM^Z zdg_Q9OKIXm)!b+(!(-lK2Vm_=F>uhLi_&|##CfNf4;Ptc7dn(e0w*adhS#hvWtPXo zYU3@2zefw)v7-TNFnDWefA?3X2I>_IDfdhEH;cOrBnR$Zjwjr%X`(uFg&9%fw_P2p z#+aStI8Kl_m>E23#;euL-G*M$0R-ICMVVX`pD~7Tq&{jg zOuzC`m?%Qm#(GE2rmnpDoJHKlRTFv1>cB`%ToX)4-dJ>O(!DNiMN!L7Y~}AheutUv?a8;-i~R$WvEkV~>b?E*Y$$;<|c zL2@MNVn6tYESp&Hm%)Ox=QY>XclEnksQ4t?bcx*>(FEzbHJ`&GJoqa@DGUETBc8SX zYA2Ol#cbCm1ot{i(I*+qydF9g zgNkaElYq1yN#MB=+#_jkM?M<*;P_P4dTXqPtTf=sgA6O0^mfWI5y#RkFKmg}aR65o z|A*DIyt4^b9qlM65P$3GdFDTlx4yLf2Os7t+2iZT~7zGQe=IO;*r+w}IPY6Q*B?OQoDm4omlrOpS zysZq>0KVJP6S#dBozDH6R`alDcStdyJJndm5I*qJD%4i+;a`xmwfI7F|@; zwV8tC(BGJX6VY5(`z#0&wtZF59liN+?A+KEtzg6CVhIDm9C_-|K_Xc-DJv_nf6%bM z;7?kdzT}xt6=j#cZc7N6K3>$YjiFwE7ke8bX^UK-FVHz^!Kt)dr^6!O?!bTK@A((w z{_jvpFz)}~jynWCiXsD=6bK0bocilOmbgPmA8-Ey`!4M&{(s(s|L3M$advAb;*38> T3hHCOJ>QM1mRBlG?>+lpNT3|L diff --git a/tutorials/notebook/mindinsight/mindinsight_image_histogram_scalar_tensor.ipynb b/tutorials/notebook/mindinsight/mindinsight_image_histogram_scalar_tensor.ipynb index 891f0489..d5ff873a 100644 --- a/tutorials/notebook/mindinsight/mindinsight_image_histogram_scalar_tensor.ipynb +++ b/tutorials/notebook/mindinsight/mindinsight_image_histogram_scalar_tensor.ipynb @@ -915,8 +915,8 @@ "source": [ "### 单独记录张量数据\n", "\n", - "1. 在AlexNet网络的`__init__`方法中初始化`ImageSummary`。\n", - "2. 在AlexNet网络的`construct`方法中使用`ImageSummary`算子记录输入图像。\n", + "1. 在AlexNet网络的`__init__`方法中初始化`TensorSummary`。\n", + "2. 在AlexNet网络的`construct`方法中使用`TensorSummary`算子记录张量数据。\n", "3. 配置`summary_collector`为:\n", "\n", "```python\n", @@ -1220,16 +1220,16 @@ "## 注意事项和规格\n", "\n", "- 在训练中使用Summary算子收集数据时,`HistogramSummary`算子会影响性能,所以请尽量少地使用。\n", + "- 出于性能上的考虑,MindInsight对比看板使用缓存机制加载训练的标量曲线数据,并进行以下限制:\n", + " - 对比看板只支持在缓存中的训练进行比较标量曲线对比。\n", + " - 缓存最多保留最新(按修改时间排列)的15个训练。\n", + " - 用户最多同时对比5个训练的标量曲线。\n", "- 为了控制内存占用,MindInsight对标签(tag)数目和步骤(step)数目进行了限制:\n", " - 每个训练看板的最大标签数量为300个标签。标量标签、图片标签、计算图标签、参数分布图(直方图)标签、张量标签的数量总和不得超过300个。特别地,每个训练看板最多有10个计算图标签、6个张量标签。当实际标签数量超过这一限制时,将依照MindInsight的处理顺序,保留最近处理的300个标签。\n", " - 每个训练看板的每个标量标签最多有1000个步骤的数据。当实际步骤的数目超过这一限制时,将对数据进行随机采样,以满足这一限制。\n", " - 每个训练看板的每个图片标签最多有10个步骤的数据。当实际步骤的数目超过这一限制时,将对数据进行随机采样,以满足这一限制。\n", " - 每个训练看板的每个参数分布图(直方图)标签最多有50个步骤的数据。当实际步骤的数目超过这一限制时,将对数据进行随机采样,以满足这一限制。\n", " - 每个训练看板的每个张量标签最多有20个步骤的数据。当实际步骤的数目超过这一限制时,将对数据进行随机采样,以满足这一限制。\n", - "- 出于性能上的考虑,MindInsight对比看板使用缓存机制加载训练的标量曲线数据,并进行以下限制:\n", - " - 对比看板只支持在缓存中的训练进行比较标量曲线对比。\n", - " - 缓存最多保留最新(按修改时间排列)的15个训练。\n", - " - 用户最多同时对比5个训练的标量曲线。\n", "- 由于`TensorSummary`会记录完整Tensor数据,数据量通常会比较大,为了控制内存占用和出于性能上的考虑,MindInsight对Tensor的大小以及返回前端展示的数值个数进行以下限制:\n", " - MindInsight最大支持加载含有1千万个数值的Tensor。\n", " - Tensor加载后,在张量可视的表格视图下,最大支持查看10万个数值,如果所选择的维度查询得到的数值超过这一限制,则无法显示。\n", @@ -1237,9 +1237,8 @@ "  1)避免使用`TensorSummary`记录较大的Tensor。\\\n", "  2)减少网络中`TensorSummary`算子的使用个数。\n", "- 功能使用完毕后,请及时清理不再需要的训练日志,以释放磁盘空间。\n", - "\n", "- 备注:估算`TensorSummary`空间使用量的方法如下:\n", - " - 一个`TensorSummary`数据的大小 = Tensor中的数值个数 * 4 bytes。假设使用`TensorSummary`记录的Tensor大小为32 * 1 * 256 * 256,则一个`TensorSummary`数据大约需要32 * 1 * 256 * 256 * 4 bytes = 8,388,608 bytes = 8MiB。又假设`SummaryCollector`的`collect_freq`设置为1,且训练了50个迭代。则记录这50组数据需要的空间约为50 * 8 MiB = 400MiB。需要注意的是,由于数据结构等因素的开销,实际使用的存储空间会略大于400MiB。\n", + " - 一个`TensorSummary`数据的大小 = Tensor中的数值个数 * 4 bytes。假设使用`TensorSummary`记录的Tensor大小为32 * 1 * 256 * 256,则一个`TensorSummary`数据大约需要32 * 1 * 256 * 256 * 4 bytes = 8,388,608 bytes = 8MiB。`TensorSummary`默认会记录20个步骤的数据,则记录这20组数据需要的空间约为20 * 8 MiB = 160MiB。需要注意的是,由于数据结构等因素的开销,实际使用的存储空间会略大于160MiB。\n", " - 当使用`TensorSummary`时,由于记录完整Tensor数据,训练日志文件较大,MindInsight需要更多时间解析训练日志文件,请耐心等待。" ] }, -- GitLab