Commits · master · Lei Wang / disometry

05 Feb, 2020 2 commits
- hmm, at finite setting disometry breaks variational principle · 1ea33441
  Lei Wang authored Feb 05, 2020
  
  1ea33441
- python hotrg_qr.py -cuda 0 -niter 6 -init hotrg -exact -0.9397153248292055 gets lower than exact ? · db2772e4
  Lei Wang authored Feb 05, 2020
  
  db2772e4
04 Feb, 2020 6 commits
- provide hotrg init to manopt, but still worse than mera · f3bcbe49
  Lei Wang authored Feb 04, 2020
  
  f3bcbe49
- initial block in hotrg_manopt.py · 28f06f9a
  Lei Wang authored Feb 04, 2020
  
  28f06f9a
- python hotrg_mctorch.py -chi 16 -float32 works · 594759cb
  Lei Wang authored Feb 04, 2020
  
  594759cb
- manopt works fine now · f475da04
  Lei Wang authored Feb 04, 2020
  
  f475da04
- work with geoopt's RiemannianAdam optimizer, but still no better than mera update · a22a9b3a
  Lei Wang authored Feb 04, 2020
  
  a22a9b3a
- added QR and SVD reparametrization, they do not perform better than mera update · 44f67725
  Lei Wang authored Feb 04, 2020
  
  44f67725
21 Jan, 2020 1 commit
- print truncation error, and seperate get_isometry and renormalize to handle checkpoint correctly · 23f5dacc
  Lei Wang authored Jan 21, 2020
  
  23f5dacc
02 Jan, 2020 1 commit
- better starting point in CTMRG, symmetrize T · 1d4363f0
  Lei Wang authored Jan 02, 2020
  
  1d4363f0
01 Jan, 2020 13 commits
- add checkpoint · bf55d87c
  Lei Wang authored Jan 01, 2020
  
  bf55d87c
- correct horrible mistakes in renormalize · 130d6bec
  Lei Wang authored Jan 01, 2020
  
  130d6bec
- start from random C and E to prevent it from converging to early · 2912b26a
  Lei Wang authored Jan 01, 2020
  
  2912b26a
- trying to make scaling factors right · c2edca82
  Lei Wang authored Jan 01, 2020
  
  c2edca82
- all isometry start from random · 2cbff571
  Lei Wang authored Jan 01, 2020
  
  2cbff571
- double layer · 8b448682
  Lei Wang authored Jan 01, 2020
  
  8b448682
- some initial steps of xtrg, but ctmrg still converges too fast · cb8d30a0
  Lei Wang authored Jan 01, 2020
  
  cb8d30a0
- does normalization matter ? · 3bcc50e2
  Lei Wang authored Jan 01, 2020
  
  3bcc50e2
- still not right, show we make Ux and Uy? · 7ab0db6d
  Lei Wang authored Jan 01, 2020
  
  7ab0db6d
- added ising tensor · 60be88db
  Lei Wang authored Jan 01, 2020
  
  60be88db
- adding symmetrize function, result not right · 9c69ec81
  Lei Wang authored Jan 01, 2020
  
  9c69ec81
- Merge branch 'master' of code.itp.ac.cn:wanglei/disometry · 5aa31085
  Lei Wang authored Jan 01, 2020
  
  5aa31085
- xtrg+ctmrg · 6965051c
  Lei Wang authored Jan 01, 2020
  
  6965051c
28 Dec, 2019 2 commits
- split Ux Uy · 8e0027d8
  Lei Wang authored Dec 28, 2019
  
  8e0027d8
- added isotrg_sweep.py · cfd6ad61
  Lei Wang authored Dec 28, 2019
  
  cfd6ad61
30 Sep, 2019 1 commit
- improve lbfgs in hotrg, but still performs worse then hotrg_sweep · 9a052fe5
  Lei Wang authored Sep 30, 2019
  
  9a052fe5
26 Sep, 2019 5 commits
- make the plotting script more flexible · 3b40d4df
  Lei Wang authored Sep 26, 2019
  
  3b40d4df
- can set h=0 if 1/2**niter is below the machine precision · ac14760c
  Lei Wang authored Sep 26, 2019
  
  ac14760c
- results shows diff into isometry gives bigger error in magnetization than impurity tensor method · 43763817
  Lei Wang authored Sep 26, 2019
  
  43763817
- Merge branch 'master' of gitlab.theory.iphy.ac.cn:wanglei/disometry · 236034f8
  Lei Wang authored Sep 26, 2019
  
  236034f8
- dhotrg · 2e8201e3
  Lei Wang authored Sep 26, 2019
  
  2e8201e3
19 Aug, 2019 2 commits
- torch.no_grad · fdd8dc90
  Lei Wang authored Aug 19, 2019
  
  fdd8dc90
- correct sign should be lnZ · 24f0891f
  Lei Wang authored Aug 19, 2019
  
  24f0891f
17 Aug, 2019 7 commits
- avoid long int when computing isometry dimension · 04d7f5ed
  Lei Wang authored Aug 17, 2019
  
  04d7f5ed
- added taurg_sweep · 5e7029bc
  Lei Wang authored Aug 17, 2019
  
  5e7029bc
- added a test on updates · 685e853b
  Lei Wang authored Aug 17, 2019
  
  685e853b
- more careful about isometry shapes (this is useful for the first few isometries) · 28563c09
  Lei Wang authored Aug 17, 2019
  
  28563c09
- added nstays · 555b2310
  Lei Wang authored Aug 17, 2019
  
  555b2310
- measure cuda time can be tricky · a75ddbec
  Lei Wang authored Aug 17, 2019
  
  a75ddbec
- refactor hotrg_sweep; tensordot does not seem to provide speedup (although it... · 39dd69ac
  Lei Wang authored Aug 17, 2019
```
refactor hotrg_sweep; tensordot does not seem to provide speedup (although it shows speed up in test_tensordot.py)
```
  39dd69ac