RuntimeError: Something went wrong after call_sampler.

[adacumos]

I'm fitting my data about 50 csv files but Prophet crashes and throwing runtime errors. This only happened when I switched to Ubuntu 16. On Windows, prophet runs without error.

`Optimization terminated with error:

Line search failed to achieve a sufficient decrease, no more progress can be made Traceback (most recent call last): File "get_predictions.py", line 283, in m.fit(df) File "/home/arvin/anaconda3/envs/devenv/lib/python3.5/site-packages/fbprophet-0.1.1-py3.5.egg/fbprophet/forecaster.py", line 536, in fit params = model.optimizing(dat, init=stan_init, iter=1e4, **kwargs) File "/home/user/anaconda3/envs/devenv/lib/python3.5/site-packages/pystan/model.py", line 510, in optimizing ret, sample = fit._call_sampler(stan_args) File "stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877.pyx", line 814, in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877.StanFit4Model._call_sampler (/tmp/tmp6in6jxib/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877.cpp:16904) File "stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877.pyx", line 401, in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877._call_sampler (/tmp/tmp6in6jxib/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8714900159976324877.cpp:8973) RuntimeError: Something went wrong after call_sampler. ` TIA

[adacumos]

This is my data daily close of a stock price.

ds,y 04/01/2016,0.072 04/04/2016,0.069 04/05/2016,0.061 04/06/2016,0.063 04/07/2016,0.062 04/08/2016,0.062 04/11/2016,0.061 04/12/2016,0.065 04/13/2016,0.064 04/14/2016,0.065 04/15/2016,0.064 04/18/2016,0.062 04/19/2016,0.062 04/20/2016,0.062 04/21/2016,0.059 04/22/2016,0.058 04/25/2016,0.057 04/26/2016,0.057 04/27/2016,0.054 04/28/2016,0.054 04/29/2016,0.058 05/02/2016,0.055 05/03/2016,0.055 05/04/2016,0.055 05/05/2016,0.055 05/06/2016,0.055 05/10/2016,0.054 05/11/2016,0.057 05/12/2016,0.058 05/13/2016,0.058 05/16/2016,0.059 05/17/2016,0.059 05/18/2016,0.058 05/19/2016,0.058 05/20/2016,0.055 05/23/2016,0.054 05/24/2016,0.054 05/25/2016,0.052 05/26/2016,0.052 05/27/2016,0.053 05/30/2016,0.053 05/31/2016,0.058 06/01/2016,0.058 06/02/2016,0.058 06/03/2016,0.058 06/06/2016,0.059 06/07/2016,0.061 06/08/2016,0.061 06/09/2016,0.059 06/10/2016,0.058 06/13/2016,0.057 06/14/2016,0.056 06/15/2016,0.058 06/16/2016,0.059 06/17/2016,0.058 06/20/2016,0.059 06/21/2016,0.061 06/22/2016,0.062 06/23/2016,0.062 06/24/2016,0.059 06/27/2016,0.059 06/28/2016,0.059 06/29/2016,0.059 06/30/2016,0.058 07/01/2016,0.058 07/04/2016,0.059 07/05/2016,0.058 07/07/2016,0.058 07/08/2016,0.058 07/11/2016,0.06 07/12/2016,0.06 07/13/2016,0.059 07/14/2016,0.088 07/15/2016,0.099 07/18/2016,0.094 07/19/2016,0.102 07/20/2016,0.107 07/21/2016,0.097 07/22/2016,0.099 07/25/2016,0.099 07/26/2016,0.099 07/27/2016,0.1 07/28/2016,0.118 07/29/2016,0.114 08/01/2016,0.109 08/02/2016,0.108 08/03/2016,0.105 08/04/2016,0.103 08/05/2016,0.103 08/08/2016,0.103 08/09/2016,0.101 08/10/2016,0.103 08/11/2016,0.095 08/12/2016,0.093 08/15/2016,0.095 08/16/2016,0.099 08/17/2016,0.097 08/18/2016,0.095 08/19/2016,0.095 08/22/2016,0.094 08/23/2016,0.097 08/24/2016,0.093 08/25/2016,0.093 08/26/2016,0.1 08/30/2016,0.098 08/31/2016,0.095 09/01/2016,0.093 09/02/2016,0.094 09/05/2016,0.098 09/06/2016,0.098 09/07/2016,0.107 09/08/2016,0.1 09/09/2016,0.099 09/13/2016,0.093 09/14/2016,0.094 09/15/2016,0.096 09/16/2016,0.101 09/19/2016,0.1 09/20/2016,0.096 09/21/2016,0.099 09/22/2016,0.1 09/23/2016,0.102 09/26/2016,0.098 09/27/2016,0.095 09/28/2016,0.099 09/29/2016,0.098 09/30/2016,0.097 10/03/2016,0.094 10/04/2016,0.094 10/05/2016,0.094 10/06/2016,0.094 10/07/2016,0.094 10/10/2016,0.094 10/11/2016,0.094 10/12/2016,0.09 10/13/2016,0.083 10/14/2016,0.089 10/17/2016,0.087 10/18/2016,0.085 10/19/2016,0.086 10/20/2016,0.085 10/21/2016,0.084 10/24/2016,0.084 10/25/2016,0.085 10/26/2016,0.084 10/27/2016,0.084 10/28/2016,0.083 11/02/2016,0.083 11/03/2016,0.083 11/04/2016,0.084 11/07/2016,0.082 11/08/2016,0.082 11/09/2016,0.078 11/10/2016,0.078 11/11/2016,0.075 11/14/2016,0.073 11/15/2016,0.072 11/16/2016,0.073 11/17/2016,0.073 11/18/2016,0.077 11/21/2016,0.075 11/22/2016,0.072 11/23/2016,0.07 11/24/2016,0.071 11/25/2016,0.07 11/28/2016,0.068 11/29/2016,0.071 12/01/2016,0.068 12/02/2016,0.07 12/05/2016,0.069 12/06/2016,0.071 12/07/2016,0.069 12/08/2016,0.068 12/09/2016,0.07 12/12/2016,0.074 12/13/2016,0.072 12/14/2016,0.072 12/15/2016,0.07 12/16/2016,0.07 12/19/2016,0.066 12/20/2016,0.065 12/21/2016,0.065 12/22/2016,0.063 12/23/2016,0.062 12/27/2016,0.062 12/28/2016,0.062 12/29/2016,0.063 01/03/2017,0.064 01/04/2017,0.068 01/05/2017,0.069 01/06/2017,0.069 01/09/2017,0.069 01/10/2017,0.071 01/11/2017,0.07 01/12/2017,0.07 01/13/2017,0.072 01/16/2017,0.073 01/17/2017,0.074 01/18/2017,0.07 01/19/2017,0.071 01/20/2017,0.072 01/23/2017,0.071 01/24/2017,0.072 01/25/2017,0.072 01/26/2017,0.072 01/27/2017,0.072 01/30/2017,0.07 01/31/2017,0.071 02/01/2017,0.069 02/02/2017,0.068 02/03/2017,0.067 02/06/2017,0.065 02/07/2017,0.065 02/08/2017,0.066 02/09/2017,0.066 02/10/2017,0.065 02/13/2017,0.064 02/14/2017,0.065 02/15/2017,0.065 02/16/2017,0.065 02/17/2017,0.066 02/20/2017,0.067 02/21/2017,0.066 02/22/2017,0.065 02/23/2017,0.066 02/24/2017,0.066 03/02/2017,0.066 03/03/2017,0.065 03/06/2017,0.065 03/07/2017,0.064 03/08/2017,0.064 03/09/2017,0.065 03/10/2017,0.067 03/13/2017,0.065 03/14/2017,0.064 03/15/2017,0.064 03/16/2017,0.072 03/17/2017,0.074 03/20/2017,0.075 03/21/2017,0.075 03/22/2017,0.073 03/23/2017,0.073 03/24/2017,0.074 03/27/2017,0.079 03/28/2017,0.078 03/29/2017,0.08 03/30/2017,0.078 03/31/2017,0.074 04/03/2017,0.071 04/04/2017,0.074 04/05/2017,0.072 04/06/2017,0.071 04/07/2017,0.073 04/10/2017,0.07 04/11/2017,0.07 04/12/2017,0.071 04/17/2017,0.072 04/18/2017,0.075 04/19/2017,0.074 04/20/2017,0.072 04/21/2017,0.072 04/24/2017,0.088 04/25/2017,0.085 04/26/2017,0.087 04/27/2017,0.093 05/02/2017,0.092 05/03/2017,0.091 05/04/2017,0.092 05/05/2017,0.093 05/08/2017,0.09 05/09/2017,0.086 05/09/2017,0.086 05/10/2017,0.089 05/11/2017,0.089 05/12/2017,0.089 05/15/2017,0.086 05/16/2017,0.087 05/17/2017,0.087 05/18/2017,0.085 05/19/2017,0.087 05/22/2017,0.087

[fountainheadpro]

I'm getting the same error on Mac: Disabling weekly seasonality. Run prophet with weekly_seasonality=True to override this.

RuntimeError Traceback (most recent call last)

in () 3 ts=pd.DataFrame(data={'ds': store_info.index.values, 'y': store_info['trade']}) 4 pr = Prophet() ----> 5 pr.fit(ts) 6 #future = pr.make_future_dataframe(periods=12) 7 #future /Users/szelvenskiy/anaconda2/lib/python2.7/site-packages/fbprophet/forecaster.pyc in fit(self, df, **kwargs) 533 534 else: --> 535 params = model.optimizing(dat, init=stan_init, iter=1e4, **kwargs) 536 for par in params: 537 self.params[par] = params[par].reshape((1, -1)) /Users/szelvenskiy/anaconda2/lib/python2.7/site-packages/pystan/model.pyc in optimizing(self, data, seed, init, sample_file, algorithm, verbose, as_vector, **kwargs) 508 stan_args = pystan.misc._get_valid_stan_args(stan_args) 509 --> 510 ret, sample = fit._call_sampler(stan_args) 511 pars = pystan.misc._par_vector2dict(sample['par'], m_pars, p_dims) 512 if not as_vector: stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346.pyx in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346.StanFit4Model._call_sampler (/var/folders/sp/05s637fx0q3f0dm7ws_2618r0000gp/T/tmpMejYXT/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346.cpp:16904)() stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346.pyx in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346._call_sampler (/var/folders/sp/05s637fx0q3f0dm7ws_2618r0000gp/T/tmpMejYXT/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_8842012118624695346.cpp:8973)() RuntimeError: Something went wrong after call_sampler.

[fountainheadpro]

Here is the data: ds y 0 2012-05-01 1 1 2012-08-01 1 2 2012-11-01 2 3 2013-04-01 1 4 2013-06-01 1 5 2013-08-01 1 6 2014-05-01 1 7 2015-11-01 1 8 2016-06-01 1 9 2016-07-01 2 10 2016-10-01 1 11 2016-11-01 1 12 2016-12-01 2 13 2017-01-01 1 14 2017-02-01 3

[fountainheadpro]

In fact, it's not related to data. Same data is processed fine until. Now any call to .fit method fails. Even if it worked before with the same dataset.

[adacumos]

Got it fixed when I upgraded to ubuntu 16.0 weird huh

[bletham]

I'm not sure what is causing this, other than that it seems to be happening inside of Stan. Maybe the update to ubuntu 16.0 also upgraded Stan?

@fountainheadpro if this still isn't working could you see what version of pystan you're running?

[kevindu1993]

I have the same error: here is the data that I got error: "ds","y" 2017-03-10,166 2017-03-11,167 2017-03-12,168 2017-03-13,169 2017-03-14,170

but when I change the data to this, it run successfully: "ds","y" 2017-03-10,1 2017-03-11,2 2017-03-12,3 2017-03-13,4 2017-03-14,5

[kevindu1993]

here is the error: File "./prophet.py", line 10, in m.fit(df); File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/fbprophet/forecaster.py", line 535, in fit params = model.optimizing(dat, init=stan_init, iter=1e4, **kwargs) File "/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/pystan/model.py", line 510, in optimizing ret, sample = fit._call_sampler(stan_args) File "stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621.pyx", line 814, in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621.StanFit4Model._call_sampler (/var/folders/_3/fj7g23z935150t_q4s8xdsz00000gn/T/tmp_u6cg9mb/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621.cpp:16904) File "stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621.pyx", line 401, in stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621._call_sampler (/var/folders/_3/fj7g23z935150t_q4s8xdsz00000gn/T/tmp_u6cg9mb/stanfit4anon_model_7a788197ac3493030b72020b9ffdbe7d_3890433898330113621.cpp:8973) RuntimeError: Something went wrong after call_sampler.

[AlexBanksWatson]

Working for some but not all of my test series. Not sure why some sets of values seem to fail...

Input: example.txt

Error:

---

RuntimeError Traceback (most recent call last)

in () 1 # Fit data ----> 2 p.fit(data) F:\Anaconda3\lib\site-packages\fbprophet\forecaster.py in fit(self, df, **kwargs) 533 534 else: --> 535 params = model.optimizing(dat, init=stan_init, iter=1e4, **kwargs) 536 for par in params: 537 self.params[par] = params[par].reshape((1, -1)) F:\Anaconda3\lib\site-packages\pystan\model.py in optimizing(self, data, seed, init, sample_file, algorithm, verbose, as_vector, **kwargs) 508 stan_args = pystan.misc._get_valid_stan_args(stan_args) 509 --> 510 ret, sample = fit._call_sampler(stan_args) 511 pars = pystan.misc._par_vector2dict(sample['par'], m_pars, p_dims) 512 if not as_vector: stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422.pyx in stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422.StanFit4Model._call_sampler (C:\Users\ALEXAN~1\AppData\Local\Temp\tmpocoa45n9\stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422.cpp:17112)() stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422.pyx in stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422._call_sampler (C:\Users\ALEXAN~1\AppData\Local\Temp\tmpocoa45n9\stanfit4anon_model_9c941523472e65f75f241e35a8769919_8513495007735644422.cpp:8986)() RuntimeError: Something went wrong after call_sampler. Running on Python3 in Windows environment.

[bletham]

OK it seems that in some situations the LBFGS optimizer is failing to converge, and this is making pystan error out. In https://github.com/facebookincubator/prophet/commit/f7becb0942cd0a005d72ae307aefee431fa962d7, if LBFGS fails we fall back to Newton. This fixes the issue for the dataset from @arvindacumos . Could others who are running into this issue install from github and see if this commit fixes it for you?

[AlexBanksWatson]

@bletham Thanks - resolves my issue.