From 2f771ac00d89e3d7452cda654129a7cddd26ed0a Mon Sep 17 00:00:00 2001
From: HongYu <20734616+james77777778@users.noreply.github.com>
Date: Sat, 20 Jan 2024 19:16:21 +0800
Subject: [PATCH] Fix `InceptionV3` and add Codecov (#20)

* Fix `InceptionV3`

* Update README

* Add Codecov

* Update Codecov

* Add Codecov badge
---
 .github/workflows/actions.yml           |   7 ++++
 README.md                               |  11 +++++++
 docs/banner/kimm.png                    | Bin 0 -> 12709 bytes
 kimm/models/inception_v3.py             |  10 ++++--
 pyproject.toml                          |   1 +
 tools/convert_inception_v3_from_timm.py |  42 ++++++++++++++++++------
 6 files changed, 59 insertions(+), 12 deletions(-)
 create mode 100644 docs/banner/kimm.png

diff --git a/.github/workflows/actions.yml b/.github/workflows/actions.yml
index 5e20c41..c00266f 100644
--- a/.github/workflows/actions.yml
+++ b/.github/workflows/actions.yml
@@ -45,6 +45,13 @@ jobs:
       - name: Test with pytest
         run: |
           pytest
+          coverage xml -o coverage.xml
+      - name: Upload coverage reports to Codecov
+        uses: codecov/codecov-action@v3
+        with:
+          token: ${{ secrets.CODECOV_TOKEN }}
+          files: coverage.xml
+          flags: kimm,kimm-${{ matrix.backend }}
 
   format:
     name: Check the code format
diff --git a/README.md b/README.md
index ff38e11..1ae8d2b 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,16 @@
+<!-- markdownlint-disable MD033 -->
+<!-- markdownlint-disable MD041 -->
+
 # Keras Image Models
 
+<div align="center">
+<img width="50%" src="docs/banner/kimm.png" alt="KIMM">
+
+[![PyPI](https://img.shields.io/pypi/v/kimm)](https://pypi.org/project/kimm/)
+[![Contributions Welcome](https://img.shields.io/badge/contributions-welcome-brightgreen.svg?style=flat)](https://github.com/james77777778/kimm/issues)
+[![codecov](https://codecov.io/gh/james77777778/kimm/graph/badge.svg?token=eEha1SR80D)](https://codecov.io/gh/james77777778/kimm)
+</div>
+
 ## Description
 
 **K**eras **Im**age **M**odels (`kimm`) is a collection of image models, blocks and layers written in Keras 3. The goal is to offer SOTA models with pretrained weights in a user-friendly manner.
diff --git a/docs/banner/kimm.png b/docs/banner/kimm.png
new file mode 100644
index 0000000000000000000000000000000000000000..342645ea836da4fa0399ff6103d3f399b2474576
GIT binary patch
literal 12709
zcmeHtc|4SD`?u~kDyf7@mKKD;Z7&R=5~8xsVC;j*mSv1>xSL2NRJIV3Y-5bEjf|LE
zg)+7<CMKqB5@QQv8HV?oc<$$UexLV`-{<rG`~Jb@x}4{6UFUIpkK_0r$9dknW@#e0
z=hz-TK0ZOy%NMNp`2K(Zf4}431^m{?&nE&Gft#11A$)xM-|~L8r7P|i=i}S{8DRqn
zg;-qC@d)%+cJmB$hbxEs-vmbU@#z_a-*ofvg@;PJ!@Us!V7Zw_vYa%+6D()1Y4NAU
zO=GwZ;&Nm#+&a?I#v{_#<D93Qf&L!7a2-H^KRnb;I^5qcAVen|EcaWk4)8bcv5Fip
zDA?0W$LhkRts%f2Sk5Oj^rnuA3JQf%MyV?Y279Zho;!C=<xe#gH8mw*gi=UEK&V@|
zQb5QlK*ApB-!d-1Lp*{JH$xGD0n$8~Ztj7|P_Ufb9_cO7hM}Rs2={-a2ZSiQ{eFVv
zO;H9Gt>WexqN1w&C-1YQrS&$)>lk{5@NPF>Vhs=aeY1Jzhxq-bn{L4&aGQvmaIl;e
zJR}eq>;eBRfj4POh;cC7EfnslulA>!hSHy^N@{=Vsr-+B@!s+82^YgdO+EFs)jd_!
z-PJsl&bg_&DQT#=xhuK<>7}itq3NNfsqOWrr>e%;e*|q#`EN0o+yN<?YMN?Vn&(tC
zRJG4(p8fZmfB*Bp$J+!VJ$!ge(O3IN?!TY?9;>Iq)2<(4Qz^f1wzTrkj7^no4*U<Z
z{O?WW>G6-5ZX$#IenY_1Lj~>!_lF0B@{FXqWh74zolr!mAN=1YGV}}mZxaF7)A4f)
z@CM6;D|y1b+>m~ua{A_`78cS<(yITL)ocmSQ~3|s{|{>cT;QL@@G$q=v3O_;cf0u?
zXak1n{6<M&uuWi~pT4nMK$x4J%Kv-}fc>}1|4Xf#j{qkE+)3phrvq;Oaa(u*;H$xa
zQ_4}4P<(vSe5Mx+ZNi^U<*xKAU5sd-zxq8h?Ssu?+~mCjo-QhT?uz|+`NT;#krQ!U
z;Z29lgH``3m43YE?w(zLR3ACWUw!Pt>mL}u>hYaWqw|M%-ZIZHMfbA@7CLp!d!FWn
zbT32~PgskME`dWx`FZ`^bKHf6HKWD$*r(foP;y=`#_hvSUQoF#xMMGHkyhgu<vsU2
z00J(5y_4oeE51jD1m5N3ZCBo;3%gqW04{g;VR$nS9R2?j0X6)Ga?G*}IH#6o^4CmJ
zYqXRssQUq3Wy-K*l}FmJ=HX*+o>V1DXb*?4xSuWhXmNOlqV44E)!nfV;pElEkq@l3
z`3DEd#-Dr&V}fQ^av4EmRUrv&Z5wpP1{#%zpW!CNp1l&)8JP=u-bsmL5XnnX++pH_
z?tb>My$Fvwlg>q3nDtLN(%Wn7qz`QqCX-Lg+5=uoL*xs)T`5X6em&qCzon5HzZCHB
z2S(HC2$V>*ix@k1Y0c7HVpzE`G|@PVT3+K<*;MUUKN1}DvtN=PIP;ZWbX;M{D~BD+
zco<yP)R<A8WUw*+rQ?-fX`b%%S;u`-HT5@a*ItfC=lC6}L?iRa%ZVU*Bx^y2lBvip
zsfn>%h_EwIdwWF&B*JfpT^q`W4^?s<GIB$$7B|{imEJ?*fu2LLTzfF5Y2C72zQl8=
zaviJN^PJTu!m3?gxC+~6lew`XWlts^N&8k1t0Y_5oDJVCH5u1cyli|O8pEa4Og>Ni
z)ZiSYG|7!mK9&5gnz3eq%KJ#WH@s3PB}le3sw_aVoYvaI1(Akx&vo?Y?v^jz5#+ji
z2Q@AZ$Iu`WIw@YT(-$IBHr8i`hF>}jGS;WU;V`>}hW@dcM3+tjGTOb=*SKB1=cS6Y
z$NHhY71EG0cMQ8emZjgju}X&ao;dg+Xrk=fVO+Hn;$c3M5MJt`_fuNVR0l;{ND>&7
z$$UAFc*qJq9`T)kDOy})iJ$U9eR@=4V|`(&a&ftW(NVp9|6Q?qRwc^lup>cZyLcg*
z`>n*FtJ<F!R5o+5zWw3lIE}=$wfUg0#q-jT_qSb(?h6dQmunQN-4$wTC(g+z`Zlu>
z6OCa~(mmU};GsUQ0%ftbAm<1f%HoK2e$u#lMIf9r78^D;&|uu*y8WvQm&*H;X)F9g
zHJ0Ogx32DcGvu8@J@s?uYm<?y6ZQ!?E!9gu=OfN|xEvb9=T6mBbw0VU(FWM+x(SkH
zo}(|`15R+Y-~%w+p$-|8*N})QIj7ZfBUQ27#d_ZwgMR)+;LoU;sqS+WJ}ViBe}YS1
zgn->U<RL4s4IQ&{7NrZH+r+(xQm>LY-D<|3s^V6JNw62<TX00+%M()?1x33nNgY($
zX;P0Htvz;K4x_p40L~`sv8$kry}^2wV=p$avL;41|Gw*x8|^w4VU#m;E$4WoPgN7A
zE$|A4Nwan_B93T{XcQ;~Y<x(u8*F0Ta<$k4d~Wec2vgI(>ia~z*r(jvs^2?0v_)Fx
zC>0t7)1R-XTKJR|nyYltLo+@mT_<4s;t~TgAx|!_LRJfe8X`GmRQ=r;O?lp@U|iF^
z@mCJ5CF4?@#`ovLP)H+TU)cM@J1KY>rDs?DXGM-3*B;TJ<e>+t{ks^=?N4@;JN>Zf
zkHe+}xWAkxxuaJl`3GN|Cs@fKTNEf%L`lWuTh2QwhaDhw5{Xq6o3b<|OkUQs_^A>-
zxOh~@D|cx2o2F=sLWw{<UAal4V4`qg%v_}c7m~ldCHpYoTrWTnWH6Ia)Ng{9jQu2d
zh$Abxnp|2Ykl_+Gm6hg@-8<-l9ygDO>ONLqwI?HYe)(H}c3G{y4ZC{Lknn6fPl5u(
z_&xQFP$R5N^MGJb<BAQi8o4@av)h%s+PbuRmM&&%f=fO8Ft*A|N!rEfvb8Tqg3YV)
zYL0KM<RWh+<0X}oFO9x;>`#5VIvsMmXqP7af=1X-i7X`SBRg~~cSU`%?E!gq<!2#m
z#Qnu?E+om6U}Yh!;6-v~FBfLII74a8TUs`?Aan$?t5TToyo>l|u0QRAJX$l&X@!~h
z*tIBIwswSG*{l;tdE)5!;5xy|TzHV3&o3Iv*J75QwnMSlY1}0>r*Wpcy1to;wwAEM
zuT@|DJ?zGvOG522zgp9$G$^W<S`;$D=^6zcm0B;Si0)i3@By6ZE!QG51q{5eaKB2i
z(@0EXYI?-G^vi%jmhO`oyO1Vk2qWKiP2agAFU~;Ly^E(p?9yeAhv+IgXCu(<x}Ua>
z=1h#T&l6;KV$v1l`7>1PNiyr-q<Sdl5}igO%XKTgPX0Jye^RDp?wdn$khR3Z-+ogn
zUWnJK`t_E~!F@@K=<0B4Yg=X1JHDnX7~Az6*+-U27Cvb274_@QKUl*J<6E*l#Bg(+
zR5|@enL(lrq;c7$wX;aS`NxWn9=k%<q@q9!&zfmXW{bVH>;vR%U3?D^(!FkU_}k!7
z{TucYl+%#KDDa2RExSiz>qq1fV}ssTX0(536bwVv@i9%<uh#5~h5x)t6j$)wtWni>
z@WyDcjMX7f<0a+hvy^UKmCCE{K1b1D-z=ntyeHy>?^j`ShI;F;COhN11C}u&*4H@v
z8R+g>^!uQ85!RrV^-4|tIRRVU6w~SfsgJ|uBVi%xcqS>Gg){c|%lGyo{JkA8VAUQw
zzlLhg&t#RNVvVS}y9+k=F-T@QlsMDeT7Bn*0Ar;)s93<lrg^c>{rJQLc1X&1<{`Q|
z1F?&u?;Git<~iOlmd(1JSC$r+98}!Lv2aZ8J^f)(9sQyA_Ls)DA}!fF0K@~ZF=ApX
z_A^ONLGj`QQ5pM_7h6b+4%+9}W)r42Dv~B9Q~`Yy?M4mDQ)1TV)2FOO@>!KV)v1;f
zk~H`s(7nIB(8efS{X8RjKD<AUR9uHnbdlOPW4_R`yAg`R2VEpQ-pyN%fL(;)YL@p8
z_{1su#bW|VqB8V_jptrhtPX<&-sX&^!Np){G<nZi>f#X(bK&7J#0bg#Db`C1mnXGO
zDD-v8G$qtpHw`O%&r0;L6y}V#fEUlSm6<*#&xzEN9et$YH^W=oX-G|nZNx}o(~sAC
zNA5gK%usxUb!}oQLBl@ZYMm?2mbHnfYf6y1U91y899s4!q#YY@4tl)J^cGjSRil8N
zmCv!M>cXBL{u(lHJ-|*qG*@CX;2lkLP4CClbfk0ZMP1A<Pku_2d~1CeG@|irs3l3e
zyPD0K*Yd$8HrcIaCArONp38JX6dURGWzvfkv<|6uO|$Q@)C2}&*Gw{+ru^|0R#%0w
z6?G3judgPuXNAZkb<<{#Ya@~50Yyx@9?vj_gbsCk@7(lE57CCch){jqrRPjDz}q%B
zmC8F`Kzxb46X`%q(>Gsq{q=C+nbkdjc__LFAELoUK(O13jSsWTz;6N*Br*&vd=6C{
zX&M8!4wVK=us0Tmnto)1R?}VHS2rzwRW>T!0hk9a1|;uXbeQuU{a|>4*QL{Hg9R$L
zo$lG}G*{O>_GFPkhWzN&D7g2wy4G6aaa`22{2&=0aA<E)(&*X`4KKp!Kd@x0JWlK?
z`f~Q{Tj6$9YXq0sFh&ZtM_)fvvODEmTb(LqrzXH0syi`21S{)&d?fWD{8+8E2uSdg
z9l2e~Wk}YR{i{DNF~;FM9eHDKK7)-1MSV+3pRy7W^RJaZaXl(_p>X6I=>eu_{wJq#
zEqH0ZH><-j;HXJ<?=a_)$$j>c5IOd-w0~8+<F}P*)G5d{0U-KNe#dlxEO3+yo$jUW
zG*5hHTi=6Lu5^6&x%_AW{3?L8pQ-tvwi0>p%diUZa3mUt)Cp2~8~X7Y1snl%ac#yd
zdPXsWEDtUPe;#PVift|A0S3-EWP#V`j<2D;A1OvH4pn!gUxpGw_X}QIow5g^=)en3
zOg(9uL>LKTmPQ?MQLtR+NO=lzsEvA&=Ba=wYGt_hTa5fDT-26-A;8YETFGvfPV2ud
zwI+?($)j1-ei(Rg(MHCZj}8g?!AzlGDNgSkh4+#x$u?Tx{A(sY*|QJSGt_@)NxtEQ
z!`sOE_wVg;9)Xrc7M+9~+E(|d`3Nj5YZaz|FuGSbv3OhRFv~f>V|AtfYU7ii$FA$1
zfLz<oQ`2b(yn)l{J=fhv_9!WHvJEZ*`>X}FQ55V?B3r6iQ)~0{ocLmaI_u@<^pbIP
zPxT7B670e=!Kdw#tl@}Mi@w7E>=GpGaI291xW%ILM$vz7>q5@;-_CUUIt(cb+OhW-
zkn7>l9pxZ<Exw+U_K_8Xm<Jc}7bY-p7G!cLjyXr8nj?&qq+QxH3euSc33DtwR(Tyi
z=CF)2w7wm91F*Xf(PXX^dVIA}wM1YlWN1MHQynT1`fwG5RpuAf_{+7ZLydp%KyXp=
zWUp~M5X+D)bM(e>OAy&POcD%1W$kh)PlyZ3M`P)4HnQrNndGPSk6JZlfb~HTMqOWj
z^5>QfY_HA)Xwd7Ai+UETD;&0E9$TmNj~2`>fb>9XUDKHpHThk-Pnv^^V6~d*fw!wh
zJ@y656iY+Gx)5UMa*ubPSGxk0U9x*8+{Y&$#`oMf_1<IxQ^cBmXnbGQC#Ni$nGT`{
zwKGkH9U`wGKZp^FNy~@OEjw52{v^il-h$R;9*-C^$VVxonG;m|5!elnG)2{&Eqij=
z@*B}d`{Tj{4LGHW9*&W-L*66FV_L4o3J{|jZHKoEEtYj&S#5<)2E4e>>zi`ZK)T9l
zaH;-fU*Z7}KaUw)6tKI}P;F_R=&9*re(efgRDyxOVo1Po*)vPcne|cDBN<9OI0ewd
zUgr<><|4JCnC+SuTdI9mz9x<^;<|0t>0~AQ>}$wAqVQvn@(^1vNc6G)#$G<mP9+{@
z`CE<|;cF`&N?ug;YB*$l5Y#xm{!3~;G8xF={zakJU=kgDMH;DfDfkZnzq(q1s6yF-
zuwP$zK>OnCb_@JEX!4!ypXyW8%vIUOFIvm9>Z_f;{0)&`;z4u|A=<%YDE{r?9px0`
z5Q>3Iu||QFtHV3o&Muu{QGJ7z7v@>xNao98SxC$Sn=!dOpX3kKK%P8AUXZcJ%ibey
zj79}!{!z~oZo!E1cx(yJq_6kRcfY7~LJT_#Z70$T+Cld2doWBFp|7`g&#|7B#e_9q
zjdriPU1^{j0%Q?2tuDW{egl#GTjTATwg7mbNZJP(iz77w=<8@L#wY&@H@3f-(1e3>
z<?02(vz-9z{viV~ewXgG->xE<<!bd%Q5C?P@66MXInted%?!@q*Z#N@ti+B!c4lmr
zN;J_|6W%gWZCJ9QD`6%)QPWfkcm&Kpx>Gd*MuFFF)c2I-WjSQT?%3<V^C?M;W}Fjv
zJv_s|J~!w7)r<JIHajR+g>Re>jC^~As!51syZ6TtRr;*2HmKDV?M60n+-;M6?_-$l
zW@Q7#CY{UvuRsW+)HlBPASS|G2>U5&8X65sYZci<{ZR)(bc2<*h%sI=<}8IcsuTOy
zO$7k*Cn4+Xf^EyqtM@C*$_u&csrgQwH3HlPS9+oAT9)9{jGd(rC&&Uyn9L^j#QeO%
ztf*JkvM7c2RS-L>#auwd5Lg76J2k;M1?l|*2tmes?4bBbRpEkQTk1VT((S<urEjp=
z2ao7l3xgWh(;TWegMyyx=ftvPYjr8g<X4a%B&`nxCfgCjy6CQAWk^i+%?)m16%^nr
z=>0U@dlocP(vNPmW7E(_%?K0UI}Bp)I*kY7Y8qI|^`_r0i5JP+yPIK{B?9RVzjm|X
zmclm({m3(Ay&s6P&?qz%&H9>+bzSkVASu{Rkn#AWDK1q->7<2E&X>`f#;QMT;nXTO
z<BTbyBb$BK)o?G6xc8j^0%^7GT~debN|OfVS4RCydb2g)y#U$$z!yBz)JmgL`(t8%
zWX(%NbJ35I_4?9Jcz&5TSTjbSD>(@XgVBi~%2U~aD8!)O>Qv37w}B!N0XxTLUac(C
zmBQ>41bpm}Bf-7PDV@=?e?&CDBV(WKmFtAbV`+b`j0FHiKt;q+68f+Um@qU9B(>68
zR<Es`mz|P*J1|r@Z8dle;Bg>2qGY95U@B7YOPqU-M$v9|LusndCZ1%9U9*7F0a+|7
z0F;NLGg;o%eu+wR`?WRSXJ@;vh^yYH`G`ln9zo>?W?$Q$i#PBOC@)k7`>>a%QHkP-
zgRKB~=jaYJrbmj1izBTq(M;Ikc&%8b)C^W_qIdo(<JbQB&;qQ|0g%r&ARaxzFngR@
zA1ZEeLm^|K{T389uKLLp3qrKUT_3xSSbrF7ELV16>7vf`*`?7Fd-LfX+A=LxYljo(
zdr*}E7Q!)H*1S0{nAUh@gXihB_~`nPK_~sKwPd&=^2Kvih)zSkl<hP`33I!>Sq%hW
z9lGQ&#8M9<Lw2nj^+S8PbdA((n@>6t3cG6{(xPTaeH}D%9V${uI(y@ek2AuEB}A)G
zB$}jNsP7`2*u<;q4vHi}FV1OYW*nubv!TC=`*bvF60pRYMnRr>>6l;!%hukd)ugO7
z;)Ot+(pw^(v^!Bdy=Ma51D<j;y>gMz;m>_{s+C5ot4jYivmOwGFe>5%keiOMu+`(?
zE6l~+)?2PovuCiZgtW~lx34G@uV>QVutpxATUs0)KI+hjD4#clg9m}a+r((DD^S$M
z#hwvSEPxECqYT%za<serp2kE*H2?u|kP-bueHvMScumy!gnP1FCPjy5f}*%=_ng6W
z){h)|rT<n$v1{r3d)Azs2Df*rnqwo!<oDW&olvZ7Crx0ZMSVY1J!nY-D&to5?laBp
zYA*!*T~1eQ%rhGC5*p`ypAqAGkaGqyEo%3!A7`wNPQ0YHS|eW9lhGeF3ShB^JjhDf
zr#Z{t=rYH)La;4C0~tFSq39y6V9)F>ub3}$NI`zRH;!xPIUmBv|Kmp)i_)c|Y8F0w
z^HCpLisG_6;K=o*SX8jcF&a=_G6p&J;snlEsjr7(;^0L=uyxr{o1!r8S^IR;1gloF
zqxD6%4sN2Y83S+czgoE#ao1@jcOpq!R5z~}x_4tOEScsU;)~Ka`v46d1k#d;TFBB!
zYsdg~C?<9g`bMyM@#u#)30xAO&bW@h0cC%Um~NP^OSBXYgKqy>UxoMW%BxS`y}6?C
zQ;-;M>uDecq~|f%6!O!OBgVSV<|1*sr9!O=zzJ=yBz-3^B9Dv9B88^@Fc^^b-2Q^N
zca~9u7bN*5pVMralS-*w8fDiE*#^~VxQ)bE*Elms(+dZJ$Pz$lGJ%`YncOyN!4=%h
zxP7Qyl*}T4k+e9ZNIW`>s3M70I369=DZsKJj}NhCMI~s>SD=Ko__oIBiM$y*ybG);
z0dUF)4TA4qfeI;9dvTkvgB|h$fFg}t9<0p|)cw-yi}O$MonxAbcinh1^pJCM6Z18l
zuJOl_!9}5oqxL7Csa@VPy0QM1wHBMxJ%x&HKb{&WuG$90$?wfb@rs>(RTG{^igbK%
z=Q6;t3qj{C#tp&lGPTD6?@KJt>n}Z}VX>SP7+2xHw>AHIDjyU*(RVwnV*I*kLB9GH
z5CJ08^XJC`#}Ny>%`RlHtwg}eapcSw)aA&t<zC_g1xZ0*QB1na-r+yq##UY|63~3c
z2r~KAH1Tz#KcAGaBVlw(>tf4X(WM#cX-#H+JvgIyrmliD^)%#D)RRd~AnXwWyP6Xg
zRJVA>HH@vCy+i%y*v2JYIc~kq<3M!zGO3ph8H_+k$>g?ogJs_v5$>-Z9N%>(Sb{Zy
zU!sTfaCR%ATCUnR&*x5!o0YQ!wO{)4liqzF*g;9D%Gmu0<x=I0^!FnhFw<VDr-$Y0
zEnKUzWU4JA(RBK2*8P971~M2qaJ?&7M*iCFO8SDoE1NuW=T$z^HF9+5h;=-_UGn!!
zbcwXeLM(?q)1$9D`dPeM#XiEzqboar#a9A?h`Q8eB}bG9aM2A5Hg|(_I0^ZZuh!<G
z7Zo({o8;KcuLb*95h<D=<*Z_VCtqP;-61Bz$U1(VArbb3#ovA!%5V3J<ma$z!**tH
z@i`**sH4N*O7#jI#8$;zI;$&)HyC1N3V3T_oU(6zhd~OEiCMVuZ3E2pvF%dt3_MQI
zj&36IkTn7O-7Ko(c<|*$rxi%MZ=)A#gw&!_GM3*1bamg;-aIg@K7d<0*2{;zRXa7^
zG?#Sv*gG#%Vb1Ew@$R?9?7!AK+o<-aSnS_CUo--#F0$-sZR^6P-8*fDJ6C|r!m_sS
zDai{)>6LfFz5-h0_3z#u!YMRZWd{N@pb=Lu7_Rlw^!{lRlEm=LO%lGH#I~a~tc<yY
zcIKQX1pB#KnY}Y`f$eh#(qKrz*l`Yzvq_>umtT(sJC!IvM7>A0If#v6Ni4U{nim2V
zxYKWes?M)R_FnVP=hKcY(*Mc=<!1QJPILXs{R*MIRW=2?%E{}jS8oI?L^3E}4fY*r
znTXBpbR<j;4~&|uM-c^^9W*3PS$x#0`ASLO^}0)YM(&8U2yX65!A73HdOBxL%NHo@
zeZ`zUcOQK$srUJ>=nS<5j?pHQ^b)~gcm5#R!WuPtNbj6-pPXhTx`6TwLPaSXx>Hp$
zzaU`%M1%cgts9^Fq85T-EcZXnvk4T@KLC}c+lM&<On?Q%FO$lC<3=urS{H=VDA>WN
z2<!q{7>P)L$!Qr4Mj=gw9GrlV$txJ3Qt8$i<x(3Tk*8~vmFsU=N+b^IMQ^M|tQhH(
z&Ve!pEq(47?tPdU0l1>=>F=tVF~>p9fW}x_Eq(wCmQBWw8yeuPMR5;9{cO3$Ql+rA
zzT2Fq$3w4GfpboC=JRKkUhr@Fdr_sU?R#oSPeVdcTeke1u92hM7nCK8<^e;&KE8p9
z31CFR(E9<{(V}Ofrk(s(PlD&KWcZU1OId=zD?~sB7B3@_tGnJXHNJG~F7f(ElR(kZ
zVAOYeCMuuaS5SW8Ul~<Du~+AT)Ov5ly@!9lz4B(oQ9GK`-@3F9mclF%S+m#!G!*x?
zD7YhD+m?SKH5Z4elnBV|f%V17E*!hS%WCr+gBoJP_s2267bgPBCn~9~5eW=d<fOVa
zqQ`*RNpwjealmtzH?jh^qFKh_0%WimWBix&u_t#9XF|gw>i3&Vd~%LrugdBO9l)i3
z{JS#Y%<5Q0Vtnfrqfz9n?Wcw8o*1o^SbAq$%Wf3?xA%#GSPA;;zub?z1I`W&no6~&
z=EdmJ>q4g+xYpg9k<5&76vX0_jQQwa@+>*xOv4;1g63gPnw*>77}a5#NOe^wX@}(w
zR#&te5hf}>2ww#?qefQg$e<1G2vA*C0M*1VC*;hpx5Cp76RDC5-*f;@*B=*o|Mp^c
z&5|U*fVF#kOu{5A4*73z^OhDEj)GgMPZ2V>&9|McO&$_dGroU6UjgTkBK%mPw%ev%
z#D3qYlQkIEino1nx|0v#D&$wy-N>q+k9qX2Z`6OiA=^u<c{V;~Z$2d`x7u~(j74}4
zxHWRHu8Qj{v~xzS@rBw16r)f>Ex_hg5xM98@d$h_9rv;3%>qzc!iWK97%S{YDIi89
zXh(VT2r}6*Kt8WgSrXkBCmLm%+S;rHK68{sQn9jiI*5Cv!ui;Kzskp4cr?falHGgS
zp~l`sGAom8`Fv_EpEKZxKE4^!)XH5Ee^dEXg<1ezG9rDK&KxJJPmm<dqZH+trG{|m
zpvMVF7j0ZX?l8mPw8at9`wB044dK=AsQ`<}bxg6(Vv`IB_w)VlJ=RYZ*ZMxg02;v@
z7L-U)c0(9B#v-Y<({=vihOLG<&m5jYzwh79D}0~qFo$5T9kOC8)v_W%Qv*{%e<x-D
z%aXiT7f{@LiO7tkwHj|s6yE{LVDc=qr)#0-Lz?8kc8rG}&~|kg>PC&&;(TG^s9aO~
z72?O|%l!D_g=YcS@wwS)(n}zz^8tmIhcM_(F$bCWDj8p9w4o{H{G_tb5P(JM#l0Ct
z34yPw!q-VpAd}taO3<@)FIQ>S{hfD`(ThRMEas(q(4C5xoVd%%4g{Qm&Qz_to|lDi
z_J#7K6tI>`)XFD#knXn!L;s?BM+;q78eg%li-on@XI6-Ju&Ok~wOE(E1~-61LN2_s
z0EKAxzg+8F%6(~~l->;4=+CznH46G}AA6S<qZ8{-jhK2W1Wt2FR0Tr5NPnE!iagsJ
ze)o~=Ku?{sB@iZX+>nWYOGS!PPeH*6w|K?z1B}dCWAnl_1ljL|1SA5SA^EH3TQXi9
z_Wg<|hWemSlO8|^TR>1b=%3A(7dL7qtEh+Vt>P{&oEyNUxiCmtM27ERd6`A%l#h8<
zrn5stl1;1qOl2ti!9VH~0euM^dFI_S&OgE&nfX4ci5i7YUK2Y3q*HZf<7JN*0!ky%
zL-2^s94{@eh$o>i0C(<J2JD!=N~bL#G8bYZZ+2zB`s6>}0+o`TYX4W+UE{j4ygJrd
zEom;a4EsJP$ZrQO*j^O-@jSIQ^U*F9usP8CNgTtK_cSfFiO(LuO$7VwGEW(;wum?O
zQpnMg&{QR_bKLi_y4e%u(8b02rKoN<TQpF0@q&8Q-V8Ej72imz5L*MXmMCS$Nl52i
zmk-DY$jf0M>U?TG4i^h^!g}Sdp&#ux(A#JvW+kcjJg8q1$JnNP|FK9<MB;tF5xs>S
zZk*EqFS9G#6P31HC9LI840iT=Y9jk5;Q2ZT!p9qmnE_9h3PS<eSW!jo)?&n)%fg_k
z$-rL$j>~I!R2Om69n*uZJC1*ujBy(7i(@Rebu<bX2PB_K%;r3T97@`xN-v}#ml&{7
zF1d>KDLTYSo_}Y{jkc@8>s^eO^Yujn6~}>{k&atfmWOrPo#d+7hU4}-eBsd3W@i2J
zFmE^_r)*%~BGz^#^Ay>|XL$6DL4suai*~h@u_tv~Jw#yR%XX9rlSpM;kH@#qv5HN?
zI-`KxCUVD0KN}tmY@4Z;CR#)+zW1t3KKq)he;ORl84r>YjCa+{G+{RaduUCp;g<o6
zrnHVQdL1L;MkRT?I6AiG39updY!}cBvN{NID|<~ctf54LqAIjfpE98#Fhjy?g2tCR
znkb`OS>LwrlO`B4>MP&I{>xRf{gQuO08CX2;0EiHHM!rGzp;9*i)wKFuiU$|83KXb
zs6tCC(u+M6F$T+qqb*$?!1q#<^JVpbqFJ&4VbqxJ0{;MuhxOfdxG6Dr?TbFJFN38N
z5{D5bYd_L~y#ggq8YZmEX<P?EA1s`%Qw6q*e^aU#vXBggfzRhC$}>7R^|cQBJC8CH
z4&ZKT@CP)&>op2SVk)<p2TeSd9JpfE$xFigcIYe{lM(6|tqGr0q34iaGGxl$?JwTX
z@yGevsuxyyJ8tocfgR>)s0fvkbq=mw=hxd69FU7ht!DnwU78WN>>ZSjkUrCZS!!n+
zGm}Opz6y6xm#-4l&-`}ZmD*l{jm0WvqMe)9;Pzr@QT@3`uW-mi{_NfdK9ggQw~7_{
zJsGq00liZzF=KQ%I0G5#5L**jC!UtzloIb6Umy7HO;mZuDb?!_wCJGdkPnq;|0AXs
z2*Z@qWz^hTm9HLWmRY{w&Yj_whXv4`2At<#?}ZXuMc;J2-s+i`?a8>B{~L*<iGr46
z1L-jX5eW~0=q!FLJWX~D*miGwe*BbpZJL!0T_V!wm!`#di?HS0d?20P2e4w!e#IVi
z=avtD3~XN_8&I@{UVy<q`BmlkV?SzcZ-E^I(V`S|TcBduL(HaElICAM<l|0sJa9#(
zWcI~5$a(*|VJ190vpA|t={6v*YpyeH{59ymuDV240OH7v<m)4{uWhIN^0U;tZ(fBH
zf{8`DQ}p_Zl0AB^Z4n`$R-3hscQCP?e%6R<ueO}|S(v*+uu5`o%fuj&NW~0aFtN_=
zB?qXyl%`CR$;;WKIoAH#C#xrIf&SnP<;?GPIk7i{uPNF)DnJk;u`8LJmv%*m6Q15O
z-U7QAsBn!y(=Rh(Ercr8i|Up4Kj8X{wM-=MlB&r4aO{|G38{Lh2dQ0PkUFGM^XrRH
zy9=*b<<fGW%l{mH7)17aENigKc8b-b6ZrAkNZX?NYV^3*?wqD$&nLH<{)cx^7))4W
ztih5~l`!ihbkJ}VjS8xK3fZoLp{jUcYo?FaJfU{Eu54s@*t69uXnmT(AucblFV${S
zB}+)qpV-1rZMFvUF0mioR>owB|00Rj2S8tD=3<t=M92+l+Lpf2c^}e~L`H7MP@7;%
z7U2We73p;fcFZT5U4BkzWOM4qwS0peia0-@!}3sTk+MEu9iaMZ<dxcmp)JUl@n_5i
zPUXWhQRHv>-w$ZNY<~>EMe5jJT8vlA`#K(KRU{UXM+5H0_b`F%@`c%7cysOS!nx>s
z&C6fi0-b<k3cO}A(8^g?o=MTqrk#7yKflrDTo_ttb(y%0iM#zLhH4Oityj8%Hy6J1
zwomHQX@BGW0sMCLkh03YjH59)5-}uVkn{KU1oM9re4y~31^D6Zrxf`(eO6WNtk&T9
z1Xp&Tb(hAtU6h1>PO?^OpL+~5{f)s%ftHCo&F}ok==o8*6A9^2+{h)kZ|TCbe>ID<
zvRwsYS$Ar>gV{exlt#Cp-aK7Z9yQ{AkbW^T#LYQE+Sklncs4j8nDyHa;xn(KwlhMl
zRYhU@U7pv$+8>S^Dyna`In{vL{lkbG>C`n71{<<inHS}=_xbAncl>s-AdAv0xkQ+K
zOZBZOUW2mVB7t{6<i2ZsyE&Kot6l7y$Z$7v;d9Jt*Ls%VJ;g17wk_GnOw>joNc5`S
zmtAmEM?fzBpx;#4QMuN?hpSnieSbAC^nF>ph`q0f5lPl}8HSPrvwJ&mnom>*E$#r@
z=D#<BvaqiAAOMjeB+qC2e$xKhd>iP&iyymNd#Zj{v;<)O;2f#wSptv8UL=sxtr133
zgqikJ-)ChwujcPWG<(`cY#{=0zk>bOtX^z0H=)_MMSurQ<=8Gjc<Q>-<qiwFC`s-e
z5)>cM5N`O@_NHVv7eK{`dvVV@ptJV`xnlGpmaS(NwWIIhzmnq>7{PsdE}$}k%LdWW
zC|U$6cbg%>%hy~OhwbvDyZcnFzZSd0b1+2=Rj^2)eng8BGne}6>4|(>#hL9wFIn)P
z|28Hf(h4m+c>Mv4aN(WuOaa76GP)Oun>f-KZHzyJd$?4$Kf)O!Vx^qes^?9JbH~@+
z?9rz;DIq7x8?(H%gPCd5K@z;9O~dZqv|M-a)q@0$0t413=6uBB&^%CMLaCNf(^>Gq
zS09w{oYx7iGC}|!Wgf-h$J4B@N{|68WiBNuD-2e#+!*Wr1+{{Iow!|>{^hax{jojZ
zTA_i5z$vc%NlDx;B(2My;3vwero5)ypmBkCDpjS>PW2*?-r5Q&OMP|giJGEvhx1GL
zm-DHbzW1lZf=dG)OC*lyi8Y|OdIlSMn@6lvTlz>RB+LTG9W$E8sRazyOz6)X6tGnr
z>z|1#IuPpjvG;f)_9K6jWqVb7SX2fKItLuRTLUAH2fZa=;3p3R=YG_D9=|bv2qYdY
zvK{=leL`!Ae9Zlt{a}}1=W)Z0`T79Z(o5&4shjGMNTK)V9ifvtwE49kN`IIM2Q<~C
z9uTnefzC8|15V;i2+c*_D<|$X{C$SvC_23UQy-GgLASPDa_Z;ILedphhSka}Jracc
zbU$HNc5nDVtBK?JP#CBp{6_tare?}&Ik)!1k86oFw#s!Vp9avba=R|?mhc888Pjt6
zg!;_Z9{gBDm;2CI;L)$x`SF%^SY_i#SWb`MHCx@8S0NAvJg8q(XuesJ9YYFo2L;79
zV1ToC=$q8{EJr*0AK)WhrzDACW{e-JZ{PN^w1>|a1Uf!gBR~4ij))aM>t1idU|m4=
z3hpRSy$iMDHiCX7O4uLN#jk?`Xw^S;mcHqyyc}Bn+0^Ki!tqo38hPB<$?XJiV$(y)
z>E1>AHCoAQ(674(^Wpv}%}j^4nriXY`&DbIaJ11hdiX&m?_>V|q1^xFiRI3XZS>M}
Vb-fZ6oxCHvrWY+Q;Eh}{{|6&RWaj_?

literal 0
HcmV?d00001

diff --git a/kimm/models/inception_v3.py b/kimm/models/inception_v3.py
index 602c105..8800f53 100644
--- a/kimm/models/inception_v3.py
+++ b/kimm/models/inception_v3.py
@@ -299,8 +299,14 @@ def __init__(
     ):
         kwargs = self.fix_config(kwargs)
         if weights == "imagenet":
-            has_aux_logits = False
-            file_name = "inceptionv3_inception_v3.gluon_in1k.keras"
+            if has_aux_logits:
+                file_name = (
+                    "inceptionv3_inception_v3.gluon_in1k_aux_logits.keras"
+                )
+            else:
+                file_name = (
+                    "inceptionv3_inception_v3.gluon_in1k_no_aux_logits.keras"
+                )
             kwargs["weights_url"] = f"{self.default_origin}/{file_name}"
         super().__init__(
             has_aux_logits,
diff --git a/pyproject.toml b/pyproject.toml
index 17c6ebc..d2eed04 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -33,6 +33,7 @@ tests = [
     "black",
     "pytest",
     "pytest-cov",
+    "coverage",
 ]
 examples = ["opencv-python", "matplotlib"]
 
diff --git a/tools/convert_inception_v3_from_timm.py b/tools/convert_inception_v3_from_timm.py
index e1acc53..5129b19 100644
--- a/tools/convert_inception_v3_from_timm.py
+++ b/tools/convert_inception_v3_from_timm.py
@@ -17,30 +17,36 @@
 
 timm_model_names = [
     "inception_v3.gluon_in1k",
+    "inception_v3.gluon_in1k",
 ]
 keras_model_classes = [
     inception_v3.InceptionV3,
+    inception_v3.InceptionV3,
 ]
+has_aux_logits_list = [True, False]
 
-for timm_model_name, keras_model_class in zip(
-    timm_model_names, keras_model_classes
+for timm_model_name, keras_model_class, has_aux_logits in zip(
+    timm_model_names,
+    keras_model_classes,
+    has_aux_logits_list,
 ):
     """
     Prepare timm model and keras model
     """
     input_shape = [299, 299, 3]
     torch_model = timm.create_model(
-        timm_model_name, pretrained=True, aux_logits=False
+        timm_model_name, pretrained=True, aux_logits=has_aux_logits
     )
     torch_model = torch_model.eval()
     trainable_state_dict, non_trainable_state_dict = separate_torch_state_dict(
         torch_model.state_dict()
     )
     keras_model = keras_model_class(
-        has_aux_logits=False,
+        has_aux_logits=has_aux_logits,
         input_shape=input_shape,
         include_preprocessing=False,
         classifier_activation="linear",
+        weights=None,
     )
     trainable_weights, non_trainable_weights = separate_keras_weights(
         keras_model
@@ -129,17 +135,33 @@
     np.random.seed(2023)
     keras_data = np.random.uniform(size=[1] + input_shape).astype("float32")
     torch_data = torch.from_numpy(np.transpose(keras_data, [0, 3, 1, 2]))
-    torch_y = torch_model(torch_data)
-    keras_y = keras_model(keras_data, training=False)
-    torch_y = torch_y.detach().cpu().numpy()
-    keras_y = keras.ops.convert_to_numpy(keras_y)
-    np.testing.assert_allclose(torch_y, keras_y, atol=1e-5)
+    if has_aux_logits:
+        torch_y = torch_model(torch_data)[0]
+        keras_y = keras_model(keras_data, training=False)[0]
+        torch_y = torch_y.detach().cpu().numpy()
+        keras_y = keras.ops.convert_to_numpy(keras_y)
+        np.testing.assert_allclose(torch_y, keras_y, atol=1e-5)
+    else:
+        torch_y = torch_model(torch_data)
+        keras_y = keras_model(keras_data, training=False)
+        torch_y = torch_y.detach().cpu().numpy()
+        keras_y = keras.ops.convert_to_numpy(keras_y)
+        np.testing.assert_allclose(torch_y, keras_y, atol=1e-5)
     print(f"{keras_model_class.__name__}: output matched!")
 
     """
     Save converted model
     """
     os.makedirs("exported", exist_ok=True)
-    export_path = f"exported/{keras_model.name.lower()}_{timm_model_name}.keras"
+    if has_aux_logits:
+        export_path = (
+            f"exported/{keras_model.name.lower()}_{timm_model_name}_"
+            "aux_logits.keras"
+        )
+    else:
+        export_path = (
+            f"exported/{keras_model.name.lower()}_{timm_model_name}_"
+            "no_aux_logits.keras"
+        )
     keras_model.save(export_path)
     print(f"Export to {export_path}")