| 2023-03-08 |
Jan | randomize initial weights
|
commit | commitdiff | tree |
| 2023-03-08 |
Jan | format source only
|
commit | commitdiff | tree |
| 2023-02-27 |
Jan | Merge branch 'tracediff' into merge
|
commit | commitdiff | tree |
| 2023-02-25 |
Jan | adding print line to initialize model_fit hiddent state...
|
commit | commitdiff | tree |
| 2023-02-21 |
Jan | A seemingly inconsequential one line edit changes results...
|
commit | commitdiff | tree |
| 2023-02-21 |
Jan | bad now good, check 5: 5.55077327554663e+19
|
commit | commitdiff | tree |
| 2023-02-21 |
Jan | orig still good
|
commit | commitdiff | tree |
| 2023-02-21 |
Jan | good stil good
|
commit | commitdiff | tree |
| 2023-02-21 |
Jan | bad still bad
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | cp fmda_kf_rnn_orig.ipynb from 40ff2937c501fc325 as...
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | print arguments of create_RNN_2
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | print id on call of create_RNN_2
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | adding outputs, check5 different
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | print arguments of create_RNN_2
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | adding print_args
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | print id on call of create_RNN_2
|
commit | commitdiff | tree |
| 2023-02-20 |
Jan | model_fit=moisture_rnn.create_RNN_2
|
commit | commitdiff | tree |
| 2023-02-19 |
Jan | check 5 values same when using moisture_rnn.create_RNN_2...
|
commit | commitdiff | tree |
| 2023-02-19 |
Jan | fixing print of ytrain.shape
|
commit | commitdiff | tree |
| 2023-02-19 |
Jan | in model.fit verbose 1 -> 2
|
commit | commitdiff | tree |
| 2023-02-19 |
Jan | in fmda_kf_rnn_orig call moisture_rnn.create_RNN_2...
|
commit | commitdiff | tree |
| 2023-02-18 |
Jan | adding prints of weights
|
commit | commitdiff | tree |
| 2023-02-17 |
Jan | print weights after training to compare actual values
|
commit | commitdiff | tree |
| 2023-02-17 |
Jan | calling reproducibility.set_seed() after check 4
|
commit | commitdiff | tree |
| 2023-02-17 |
Jan | fmda_kf_rnn_orig and fmda_rnn_rain still run the same...
|
commit | commitdiff | tree |
| 2023-02-17 |
Jan | finalizing reproducibility, calling at the beginning
|
commit | commitdiff | tree |
| 2023-02-17 |
Jan | verfied fmda_kf_rnn_orig and fmda_rnn_rain run the...
|
commit | commitdiff | tree |
| 2023-01-10 |
Jan | keeping all varibles in case_data, still similar pictures...
|
commit | commitdiff | tree |
| 2023-01-10 |
Jan | fmda_kf_rnn_orig passes date to fmda_rnn_rain, but...
|
commit | commitdiff | tree |
| 2023-01-09 |
Jan | Merge branch 'baseline' into jm_rnn
|
commit | commitdiff | tree |
| 2023-01-09 |
Jan | loadinf from json
|
commit | commitdiff | tree |
| 2023-01-09 |
Jan | mv orig files to fmda
|
commit | commitdiff | tree |
| 2023-01-09 |
Jan | saving result to jsons
|
commit | commitdiff | tree |
| 2023-01-09 |
Jan | writing result to json files
|
commit | commitdiff | tree |
| 2023-01-08 |
Jan | replacing scatter by plot - baseline figs
|
commit | commitdiff | tree |
| 2023-01-08 |
Jan | passing around only dictionary
|
commit | commitdiff | tree |
| 2023-01-08 |
Jan | storing all in a single dictionary
|
commit | commitdiff | tree |
| 2023-01-07 |
Jan | adding rain to synthetic data
|
commit | commitdiff | tree |
| 2023-01-07 |
Jan | adding function plot_data for data dictionary
|
commit | commitdiff | tree |
| 2023-01-07 |
Jan | fixing synthetic_data, mv check_data before
|
commit | commitdiff | tree |
| 2023-01-07 |
Jan | check_data prints also E
|
commit | commitdiff | tree |
| 2023-01-07 |
Jan | fixing check_data
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | fmda_rnn_rain.ipynb
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | synthetic data
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | adding data_utils
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | Scaling in moisture_rnn.py is off but would not work...
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | meso_api token run out, replaced from synopticdata.com
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | plotting data as line not dots
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | cleanup - rm create_RNN seq2batches
|
commit | commitdiff | tree |
| 2023-01-06 |
Jan | splitting moisture_rnn.py off moisture_models.py
|
commit | commitdiff | tree |
| 2023-01-05 |
Jan | zoom divide editing
|
commit | commitdiff | tree |
| 2022-11-29 |
Jan | replacing ext_kf by import from moisture_models.py
|
commit | commitdiff | tree |
| 2022-11-14 |
Jan | running find_RAWS_statio, saving with outputsn
|
commit | commitdiff | tree |
| 2022-11-06 |
Jan | adding to gitignore
|
commit | commitdiff | tree |
| 2022-11-06 |
Jan | adding README.md
|
commit | commitdiff | tree |
| 2022-11-06 |
Jan | adding src directory
|
commit | commitdiff | tree |
| 2022-11-06 |
Jan | reorganized
|
commit | commitdiff | tree |
| 2022-09-02 |
Jan | for datascience symposium
|
commit | commitdiff | tree |
| 2022-07-12 |
Jan Mandel | RNN prediction better than augmented KF, still without...
|
commit | commitdiff | tree |
| 2022-07-12 |
Jan Mandel | RNN prediction better than augmented KF, still without...
|
commit | commitdiff | tree |
| 2022-07-12 |
Jan Mandel | rnn from single equilibrium, varying hidden_units
|
commit | commitdiff | tree |
| 2022-07-11 |
Jan Mandel | with first rnn forecast
|
commit | commitdiff | tree |
| 2022-07-09 |
Jan Mandel | checked that fit works with timesteps > 1
|
commit | commitdiff | tree |
| 2022-07-09 |
Jan Mandel | adding data noise
|
commit | commitdiff | tree |
| 2022-07-08 |
Jan Mandel | stateful model with batch consisting Jof one sample
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | stateful, one batch, numbers same
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | cryptic graph execution error
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | starting stateful, not turned on yet
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | return sequences = true, result same as ours
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | return_sequences=false, still same
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | still same
|
commit | commitdiff | tree |
| 2022-07-07 |
Jan | indexed loop computing sequence
|
commit | commitdiff | tree |
| 2022-07-05 |
Jan | so-so fitting with exact starting weights except wrong...
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | fitting from analytic start OK
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | trying to make weights converge better
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | compensates for delta, even if not perfect
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | reorg only
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | reorg, no change
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | somewhat better fitting
|
commit | commitdiff | tree |
| 2022-07-04 |
Jan | cleaned up exact/initial/fitted weights
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | using create_RNN_2 only
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | with hand solution as pretraining weights change only...
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | out1 and m the same
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | last term ouf output sequence the same in hand calc...
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | still simulated ok and actual rnn not
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | reorg, out still ok
|
commit | commitdiff | tree |
| 2022-07-03 |
Jan | handwritten single linear cell recursion ok, but not...
|
commit | commitdiff | tree |
| 2022-07-02 |
Jan | adding variabke number of dense layers and the simplest...
|
commit | commitdiff | tree |
| 2022-07-02 |
Jan | isolated create_synthetic_data
|
commit | commitdiff | tree |
| 2022-07-02 |
Jan | reorganize creating RNN for easier experiments
|
commit | commitdiff | tree |
| 2022-07-02 |
Jan | consolidate creating RNN further
|
commit | commitdiff | tree |
| 2022-07-02 |
Jan | consolidate creating RNN
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | adding a dense layer, still not much convergence
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | fixing plot
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | with scaling and stateful fit
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | copy weights to same model with any batch size for...
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | small fixes
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | scaling, some convergence
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | training stateful
|
commit | commitdiff | tree |
| 2022-07-01 |
Jan | separating test preprocessing for RNN
|
commit | commitdiff | tree |
