chaitjo / graph-convnet-tsp Goto Github PK

Got error at line loss.backward() in Test backward pass. Do you know the reason?

ubuntu16.04 gpu2070s， cudnn version cant match

I tried to train for 500 nodes. GPU is running out of memory (CUDA out of memory). Did you try with 500 nodes ? If yes, did you do anything to solve this ?

hi,i have a question about data format, grateful for answer
in utils/google_tsp_reader.py at line: 83

tour_nodes = [int(node) - 1 for node in line[line.index('output') + 1:-1]][:-1]

if raw data is ['1', '9', '3', '2', '8', '7', '6', '4', '5', '10']
then we got [0, 8, 2, 1, 7, 6, 5, 3, 4],

i notice that comment "Don't add final connection for tour/cycle"

so do you mean that don't connect for '10' to '1' for the example?

why we drop the node '10'?

i use ubount18.04, according to your method ,the erro is :

ERROR: Command errored out with exit status 1:
command: /home/xiong/anaconda3/bin/python -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/home/graph-convnet-tsp/data/pyconcorde/setup.py'"'"'; file='"'"'/home/graph-convnet-tsp/data/pyconcorde/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' develop --no-deps
cwd: /home/graph-convnet-tsp/data/pyconcorde/
Complete output (242 lines):
running develop
running egg_info
writing pyconcorde.egg-info/PKG-INFO
writing dependency_links to pyconcorde.egg-info/dependency_links.txt
writing requirements to pyconcorde.egg-info/requires.txt
writing top-level names to pyconcorde.egg-info/top_level.txt
reading manifest file 'pyconcorde.egg-info/SOURCES.txt'
reading manifest template 'MANIFEST.in'
writing manifest file 'pyconcorde.egg-info/SOURCES.txt'
running build_ext
./concorde/
./concorde/Makefile.in
./concorde/configure.in
./concorde/configure
./concorde/aclocal.m4
./concorde/config.guess
./concorde/config.sub
./concorde/install-sh
./concorde/BIGGUY/
./concorde/BIGGUY/bg_test.c
./concorde/BIGGUY/bigguy.c
./concorde/BIGGUY/bigguy.tst
./concorde/BIGGUY/Makefile.in
./concorde/COMBS/
./concorde/COMBS/block.c
./concorde/COMBS/co_main.c
./concorde/COMBS/dngreedy.c
./concorde/COMBS/Makefile.in
./concorde/CUT/
./concorde/CUT/connect.c
./concorde/CUT/cut_st.c
./concorde/CUT/gomoryhu.c
./concorde/CUT/lin_all.c
./concorde/CUT/mc_main.c
./concorde/CUT/mincut.c
./concorde/CUT/segments.c
./concorde/CUT/shrink.c
./concorde/CUT/Makefile.in
./concorde/EDGEGEN/
./concorde/EDGEGEN/delaunay.c
./concorde/EDGEGEN/edgegen.c
./concorde/EDGEGEN/eg_main.c
./concorde/EDGEGEN/mlinkern.c
./concorde/EDGEGEN/xnear.c
./concorde/EDGEGEN/Makefile.in
./concorde/FMATCH/
./concorde/FMATCH/fmatch.c
./concorde/FMATCH/fm_main.c
./concorde/FMATCH/Makefile.in
./concorde/HELDKARP/
./concorde/HELDKARP/heldkarp.c
./concorde/HELDKARP/hk_main.c
./concorde/HELDKARP/Makefile.in
./concorde/INCLUDE/
./concorde/INCLUDE/bigguy.h
./concorde/INCLUDE/combs.h
./concorde/INCLUDE/consec1.h
./concorde/INCLUDE/cut.h
./concorde/INCLUDE/cuttree.h
./concorde/INCLUDE/delaunay.h
./concorde/INCLUDE/edgegen.h
./concorde/INCLUDE/fmatch.h
./concorde/INCLUDE/heldkarp.h
./concorde/INCLUDE/kdtree.h
./concorde/INCLUDE/linkern.h
./concorde/INCLUDE/localcut.h
./concorde/INCLUDE/lp.h
./concorde/INCLUDE/machdefs.h
./concorde/INCLUDE/macrorus.h
./concorde/INCLUDE/mlinkern.h
./concorde/INCLUDE/necklace.h
./concorde/INCLUDE/pq.h
./concorde/INCLUDE/pqsets.h
./concorde/INCLUDE/tinytsp.h
./concorde/INCLUDE/tsp.h
./concorde/INCLUDE/util.h
./concorde/INCLUDE/verify.h
./concorde/INCLUDE/Makefile.common.in
./concorde/INCLUDE/config.h.in
./concorde/KDTREE/
./concorde/KDTREE/kdbuild.c
./concorde/KDTREE/kd_main.c
./concorde/KDTREE/kdnear.c
./concorde/KDTREE/kdspan.c
./concorde/KDTREE/kdtwoopt.c
./concorde/KDTREE/Makefile.in
./concorde/LINKERN/
./concorde/LINKERN/flip_two.c
./concorde/LINKERN/linkern.c
./concorde/LINKERN/linkern_fixed.c
./concorde/LINKERN/linkern_path.c
./concorde/LINKERN/lk_main.c
./concorde/LINKERN/testpath.c
./concorde/LINKERN/Makefile.in
./concorde/LOCALCUT/
./concorde/LOCALCUT/checker.c
./concorde/LOCALCUT/chunks.c
./concorde/LOCALCUT/first.c
./concorde/LOCALCUT/first_main.c
./concorde/LOCALCUT/intmat.c
./concorde/LOCALCUT/lift.c
./concorde/LOCALCUT/localcut.c
./concorde/LOCALCUT/loc_main.c
./concorde/LOCALCUT/peeler2.c
./concorde/LOCALCUT/peeler.c
./concorde/LOCALCUT/separate.c
./concorde/LOCALCUT/tsporacl.c
./concorde/LOCALCUT/Makefile.in
./concorde/LP/
./concorde/LP/lpcplex4.c
./concorde/LP/lpcplex5.c
./concorde/LP/lpcplex6.c
./concorde/LP/lpcplex8.c
./concorde/LP/lpnone.c
./concorde/LP/lpqsopt.c
./concorde/LP/Makefile.in
./concorde/PQ/
./concorde/PQ/consec1.c
./concorde/PQ/cuttree.c
./concorde/PQ/necklace.c
./concorde/PQ/pq.c
./concorde/PQ/pqtest.c
./concorde/PQ/Makefile.in
./concorde/TINY/
./concorde/TINY/bnbmsp.c
./concorde/TINY/tinytsp.c
./concorde/TINY/tt_main.c
./concorde/TINY/randtsp.awk
./concorde/TINY/Makefile.in
./concorde/TOOLS/
./concorde/TOOLS/edg2len.c
./concorde/TOOLS/edgunion.c
./concorde/TOOLS/fconvert.c
./concorde/TOOLS/junk.c
./concorde/TOOLS/killgrun.c
./concorde/TOOLS/lkhalone.c
./concorde/TOOLS/lkhboss.c
./concorde/TOOLS/lkhgrunt.c
./concorde/TOOLS/prob2tsp.c
./concorde/TOOLS/showres.c
./concorde/TOOLS/subby.c
./concorde/TOOLS/tdivide.c
./concorde/TOOLS/tour2edg.c
./concorde/TOOLS/tourchk.c
./concorde/TOOLS/tourlen.c
./concorde/TOOLS/ttour.c
./concorde/TOOLS/Makefile.in
./concorde/TSP/
./concorde/TSP/bcontrol.c
./concorde/TSP/blkcomb.c
./concorde/TSP/blossom.c
./concorde/TSP/bosstell.c
./concorde/TSP/branch.c
./concorde/TSP/cliqhash.c
./concorde/TSP/cliqwork.c
./concorde/TSP/combcliq.c
./concorde/TSP/concorde.c
./concorde/TSP/control.c
./concorde/TSP/cutcall.c
./concorde/TSP/cutpool.c
./concorde/TSP/cutserv.c
./concorde/TSP/ddecker.c
./concorde/TSP/domboss.c
./concorde/TSP/domgrunt.c
./concorde/TSP/domtest.c
./concorde/TSP/ex_price.c
./concorde/TSP/generate.c
./concorde/TSP/growcomb.c
./concorde/TSP/poolcat.c
./concorde/TSP/prclique.c
./concorde/TSP/prob_io.c
./concorde/TSP/probserv.c
./concorde/TSP/qsparse.c
./concorde/TSP/skeleton.c
./concorde/TSP/subboss.c
./concorde/TSP/subgate.c
./concorde/TSP/subgrunt.c
./concorde/TSP/teething.c
./concorde/TSP/test_tsp.c
./concorde/TSP/tighten.c
./concorde/TSP/tsp_call.c
./concorde/TSP/tsp_lp.c
./concorde/TSP/xtour.c
./concorde/TSP/Makefile.in
./concorde/UTIL/
./concorde/UTIL/allocrus.c
./concorde/UTIL/bgetopt.c
./concorde/UTIL/dheaps_i.c
./concorde/UTIL/edgelen.c
./concorde/UTIL/edgemap.c
./concorde/UTIL/edgeutil.c
./concorde/UTIL/eunion.c
./concorde/UTIL/fastread.c
./concorde/UTIL/genhash.c
./concorde/UTIL/getdata.c
./concorde/UTIL/priority.c
./concorde/UTIL/safe_io.c
./concorde/UTIL/signal.c
./concorde/UTIL/sortrus.c
./concorde/UTIL/subdiv.c
./concorde/UTIL/urandom.c
./concorde/UTIL/util.c
./concorde/UTIL/zeit.c
./concorde/UTIL/Makefile.in
./concorde/VERIFY/
./concorde/VERIFY/verify.c
./concorde/VERIFY/ver_main.c
./concorde/VERIFY/verlist.tst
./concorde/VERIFY/Makefile.in
./concorde/README
loading cache ./config.cache
checking host system type... configure: error: can not guess host type; you must specify one
Traceback (most recent call last):
File "", line 1, in
File "/home/graph-convnet-tsp/data/pyconcorde/setup.py", line 174, in
'build_ext': build_ext,
File "/home/xiong/anaconda3/lib/python3.6/site-packages/setuptools/init.py", line 144, in setup
return distutils.core.setup(**attrs)
File "/home/xiong/anaconda3/lib/python3.6/distutils/core.py", line 148, in setup
dist.run_commands()
File "/home/xiong/anaconda3/lib/python3.6/distutils/dist.py", line 955, in run_commands
self.run_command(cmd)
File "/home/xiong/anaconda3/lib/python3.6/distutils/dist.py", line 974, in run_command
cmd_obj.run()
File "/home/xiong/anaconda3/lib/python3.6/site-packages/setuptools/command/develop.py", line 38, in run
self.install_for_development()
File "/home/xiong/anaconda3/lib/python3.6/site-packages/setuptools/command/develop.py", line 140, in install_for_development
self.run_command('build_ext')
File "/home/xiong/anaconda3/lib/python3.6/distutils/cmd.py", line 313, in run_command
self.distribution.run_command(command)
File "/home/xiong/anaconda3/lib/python3.6/distutils/dist.py", line 974, in run_command
cmd_obj.run()
File "/home/graph-convnet-tsp/data/pyconcorde/setup.py", line 115, in run
build_concorde()
File "/home/graph-convnet-tsp/data/pyconcorde/setup.py", line 99, in build_concorde
_run(cwd, "build/concorde")
File "/home/graph-convnet-tsp/data/pyconcorde/setup.py", line 76, in _run
subprocess.check_call(cmd, shell=True, cwd=cwd)
File "/home/xiong/anaconda3/lib/python3.6/subprocess.py", line 291, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command 'CFLAGS="-fPIC -O2 -g" ./configure --prefix /home/graph-convnet-tsp/data/pyconcorde/data --with-qsopt=/home/graph-convnet-tsp/data/pyconcorde/data ' returned non-zero exit status 1.
building concorde
----------------------------------------
ERROR: Command errored out with exit status 1: /home/xiong/anaconda3/bin/python -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/home/graph-convnet-tsp/data/pyconcorde/setup.py'"'"'; file='"'"'/home/graph-convnet-tsp/data/pyconcorde/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' develop --no-deps Check the logs for full command output.

Hi,

I am getting the following error on line 96 in file Utils/beamsearch.py
On this line
self.mask = self.mask.gather(1, perm_mask)

Error is

gather_out_gpu(): expected dtype int64 for index

The issue is that perm_mask is having float values but torch.gather expects integer in the index argument

Could you please help me on this.

I used the same libraries with the version numbers recommended in the readme.

I am running for tsp10.

Hi Chaitanya,

Recently I have been studying your paper ''An Efficient Graph Convolutional Network Technique for the Travelling Salesman Problem'', and tried to understand your proposed GCN model. However I have some questions to discuss with you.

In gcn_layers.py function NodeFeatures and EdgeFeatures, I found that the features of nodes and edges are aggregated as in a fully connected graph, i.e., for every node, any other nodes are all its neighbours. I think this is because in TSP context, you set the state as a fully connected graph. Am I correct?

Further more, if in other domains, such as chemistry, where graph is not fully connected, is there any model for aggregation based on connectivity been implemented, i.e. only neighburs of V will be aggregated together?

Hi, when I directly run your code in main.ipynb, other parts are correct, but I received "RuntimeError: grad_input must be contiguous"

I think this may be caused by loss_edges function, where y = y.permute(0, 3, 1, 2) which makes y not contiguous anymore? Thanks!

when I try to run this code, but it makes errors like this, how to fix it? I just want to train it in my mac and then test the model. thanks. So how should I change the code for support it run in my macbook

Hi,

Thanks for sharing your code. I was wondering how can I test the trained model on a single example from TSPLIB dataset (file with an extension .tsp rather than .txt).

Thanks

-Sara

hi, thanks for your open-source code. I met a bug when I run the beam search code in utilis/beamsearch.py line96:
self.mask = self.mask.gather(1, perm_mask)
where 'perm_mask' should be a LongTensor type, but here its type is FloatTensor. It makes severl beam search-based functions fail to use.