From 8d0cfc5fdf3109d4b9551b3daf8e0372ba3d315c Mon Sep 17 00:00:00 2001
From: Ryan Missel <rxm7244@rit.edu>
Date: Sun, 31 Mar 2019 03:13:58 -0400
Subject: [PATCH] implemented training model

---
 hypotheses_modeling/model.ckpt      | Bin 0 -> 4443 bytes
 hypotheses_modeling/pytorch_shit.py |  23 ++++++++++-------------
 2 files changed, 10 insertions(+), 13 deletions(-)
 create mode 100644 hypotheses_modeling/model.ckpt

diff --git a/hypotheses_modeling/model.ckpt b/hypotheses_modeling/model.ckpt
new file mode 100644
index 0000000000000000000000000000000000000000..7ab304026c704dccd1c109030959810b233c7987
GIT binary patch
literal 4443
zcma)AO>7&-6{alNmg%4T^e=JKcq+Tr(xE7cqGgp1X}OkVyR>A9mRmNavK%gl+KraG
zoZVTAion)RpjhdlMS>K)^is5kw#lJAq?g94QxA39Lr(>g21Xm8K+r=A6e!@LK>KE9
zwOq;)8bXk`^WMDo=6m0J^H#U_d^>Lb=Lgdd(Mvy>dpdP{&&<cY@ViZ8O6l=fEY@^v
zW^1;o-Xe}`*cR;ph0<Y=m<D4eQHfPEutoQ5l;W_+R+~h1jhBf_d)u^6DeZ+tSGOHT
z_ZIe)df~~i7~NOsEA??3wzWj}7y3L8d@7}Vu+*??o=Ip-Q`H7GELDxq5*DXDr78$7
zWQT9K5XkVHZkWS!tD26T#++lnL^S3Oy9RC$mkg0slQ>3$Sg^LD6U<f#VSJaq+O(Mt
zDy|A{h-p|$!_8G@JDRS#y1`P-RSH4&#A2F>T^E68#j;W}BBVq%j^I-xb)>3>WiVA$
zTw>Of$gr9o0|U580wu2KdJu(@NQKG?v0sQ_TdJ;&q8tMAR}G6`M~PTTqKQY8(_P9F
zNt8*%!+`3xvx1#k5Kk)-MGLlx1EySKsbL#ZhdWnViH;M8Fwe0NKM%;#AdAypM*HRy
zcj(?W-KUiL0e-cLwPi>+-LFZT5Ph)ZrkHM7adn9>m48#~o~1D$%+{Vb?N1cxfzm;4
zr->a%70(HcYrrlO@&~$Z{F+lE4yjEU8sz*SY&aA+_Ud)6eOTv<M@o-vMZ=G+>bAY?
z(#M4C@zN143lIrJPYBVIr6;4Jrb7bpr-b;^r6;)9Wtb7Q1`!0(&j{&fOGmlXv~ew}
z(lPQmq4t!FSF+fpPm9H)ieNr~PCu{cu?~2FS_+s56o^YE#OX6748AU8%Yy7>PWJNo
zW^)QNT)t{+*enC*mjFVyJh#ep+Y0`Sohh3}wd@*<3^gHDc!{`Wk*Ts}l>=v>!kuxh
zf1xygV2kSk3%j`P3=9l_U(+Bi!bt8z%)yom@HQ|{8ITJ>NkI*0b_41O)gZ_$%7RqS
zR#t(1t@AG~pyyDvBI$bGS#o71@k)RKuS0=)4so-H_YGv%5y6@+iwe28(KJa|I5yPT
zF8UyXD%`%IYG5#Yr4GzyJk)7yx+EzgHt{NP(28N2$g&wS0TT;h_^H(lX51o(?@fiS
zd9+icjy;oQQKhorc%%{>={m3c&Y6bL72Xg+5R};@I+a9aMp|smu*%^H+KCE+{^XD?
zxK342g8;e^#wxPBMim0I>yZ#-Y_^3>RO1bfcVst|520%Ugq=MDKs&j{aQ-R)$CC(6
zk97=(b`io5BUgZQz%PDWgxw(TaMV<gTE-j$Y(Xx=y(`36(g6!$o<oeQCP8x5<;ms(
z$xQwvY=@gU<^huw319aib^$eo7QpB&Y<irxPKQ3+9`D=+Hi2q2RCnygmV+{9ER@#T
zjkYpWcB#YGmfPSU_?D<7Wv+G#ZW(A^t~*#;8*Q)ckmq(K+$9N#w&B57cL8H7z8wKb
zRJsYu4+vz>V#q|c!mU=2;f8dT#P0y}2pD0Q+z|f+=>K1EcEBdG{S)9LJYQ^ya(A;?
zNhoe2pU<O4*WH&MNusfl&fUc|#N`=Yb<m+|mNJ@#TQa>(F=!Jx)Q}f$x0$8`(}^N=
z;KF>FD3Lo-UU}E~>P=yLY-CYxhc+qVS5*z_kzp7-R28DRImlBmxY)U<!(m`+atND2
z5`o*u6W^KRT_$pw%aKBM>ZG`=cONezofRU9JcT(HmwnVJ=tUp96=lhS%G}o7tw=x?
zpbgm$2Q3nP#9stKqb{0ir#tWji!ek1d?MQsa|C+%4z?V!WVkR}bfyeN<{^n@ZHsg=
zE<D4{10zdkAW<S?)gv;#Ba#o^OH#|&;1{3jSq8rq<xrTCa3vAW@xGGcT|_w@i5;Nm
z92`oNUqL*&?(@{K1*2vXaWV;KDs(8#Gx>xfWhzizd<Fc)_!oHYoRVW)wLPl_Hb&7f
z>ck^0bZZJP-ACcMNunWPBPmS<7)`DaysWyKV>Fox^PFzi1Znt;+&z+t`ivkK%$w7}
z<RfcQQaXe##3zywxI!JQI;81{!#-+pFv7gt8Z_`IFU5i=thM%FO3pA!Nb``45fed3
zQ6s^UU~D9Dp&6K*pJntIpPW&feo@i)>j(6>m;(poD~{v_Js!+~s>ck|jl-#`o?+HH
zT#VBbyQnZxbLy>hoPJ4BO8a42bfFqy-lHewmiRPGs_eR4jMG8!$~HEd$&94O$1-D?
zkqIh?QmyIHM4^9{(Npt<SfRhr^BR4w(8K>EUl&miF`6pU;SKRHa+i)4X<EF!aR}&`
zenN*B^ppBFADPT}hGukW`?RpVPR9#9>-3DYJHhCei!>|kp1n(-FVdW+%LzZP%W3|c
zE=TuCUC!v|bvccHMVI6D1zk?qQ)27}ZUBGbmGa_-c+ZF=l1>+TN{0ZaXe#NYLcdoH
zX7e<?tUm)uyV9m#jV7(2=*M?E<LNV_X?hjThycGCe#X8+|7&zs!gx)dbB@vZA}z{u
zUcXBhiu8uQJ8a{p$LZI=WFMIPx-hv;OB}XuNOLbRdb3E&(%j-*`eKo)Td-9E*l++_
zHGoYEV5<eNkpQ;3esM#eF7)e{;2pf`!IM{oH}OUO_Rrt(y9v|Ny(SX8XWJT&UjBa|
z+$%upBPsp{$v>0&(&pjm$<3>O^3FB?cxbcw(fhffzyHu*|32|A-+4E`rT4+xCpV8D
z-0MH39m@aqADPW>{`ejLdvD%3_qXA{`Zo{6^86nN{F5iX-}>{j_rCMpcl}p>c_fd-
zYAp7k7W03!Iq6@{Eau<%`LsWFiTUdT@0@L)XyrdnAN6nF|IOJCK3w)M$A{1UYvkd3
zcP9Rj`|0}|K09^v-s}H;+i(41!#~#hd;j^r{LcUI*sJ-Z{@}d_4@dmX`E35xw|?#K
n`__Yd_b<JXd-LC8`4`^0;m_Re^=JB~a{p}I&wcc(pXL7tfZA`d

literal 0
HcmV?d00001

diff --git a/hypotheses_modeling/pytorch_shit.py b/hypotheses_modeling/pytorch_shit.py
index 7902e34..483c9bc 100644
--- a/hypotheses_modeling/pytorch_shit.py
+++ b/hypotheses_modeling/pytorch_shit.py
@@ -13,8 +13,7 @@ class Net(nn.Module):
 
     def forward(self, x):
         x = torch.sigmoid(self.fc1(x))
-        x = torch.sigmoid(self.fc2(x))
-        return x
+        return self.fc2(x)
 
 
 def get_argmax(array):
@@ -25,9 +24,7 @@ def get_argmax(array):
             max = array[i]
             index = i
 
-    one_hot = [0, 0, 0, 0]
-    one_hot[index] = 1
-    return one_hot
+    return [index]
 
 
 def get_trainset(dataset, k, n0, x_columns, y_columns):
@@ -65,7 +62,7 @@ def get_trainset(dataset, k, n0, x_columns, y_columns):
             y.append(yt)
 
     x = torch.FloatTensor(x)
-    y = np.array(y)
+    y = torch.LongTensor(y)
     return x, y
 
 
@@ -75,20 +72,19 @@ def time_series_sigmoid_classification(steps, dataset, k, n0, x_columns, y_colum
   loss = nn.CrossEntropyLoss()
 
   for step in range(steps):
-      x, y = get_trainset(dataset, k, n0, x_columns, y_columns)
+      optimizer.zero_grad()
 
+      x, y = get_trainset(dataset, k, n0, x_columns, y_columns)
       pred = net(x)
-      pred = pred.detach().numpy()
-
-      for row in range(len(pred)):
-        pred[row] = get_argmax(pred[row])
-
-      net_loss = loss(pred, y)
+      net_loss = loss(pred, torch.max(y, 1)[1])
       net_loss.backward()
       optimizer.step()
 
       print("Loss at Step {}: {}".format(step, net_loss))
 
+  x, y = get_trainset(dataset, k, n0, x_columns, y_columns)
+  accuracy(net, x, y)
+
 
 def accuracy(net, x, y):
     pred = net(x)
@@ -106,6 +102,7 @@ def accuracy(net, x, y):
 
     accuracy = (correct / total) * 100
     print("Accuracy for set: {}%".format(accuracy))
+    torch.save(net, "model.ckpt")
 
 
 def main():