Log - jacobian - a basic keras-like neural network library for c++/python

Date	Commit message	Author	Files	+	-
2026-07-25 22:59	Clean up files for easier reading, remove GSL dependency	quantumish	13	+649	-684
2021-04-10 04:23	Properly merge perf	David Freifeld	2	+8	-16
2021-04-10 04:16	Merge branch 'perf'	David Freifeld	2	+43	-10
2021-04-10 03:41	Move optimizers to util.hpp, consolidate namespaces, fix python demo	David Freifeld	8	+172	-177
2021-04-10 03:05	Remove pointer indirection	David Freifeld	5	+413	-422
2021-03-29 18:43	Remove old scripts/benchmarks	David Freifeld	7	+0	-361
2021-03-29 18:41	Update python examples to use submodules	David Freifeld	2	+15	-16
2021-03-29 18:37	Merge branch 'master' of https://github.com/richardfeynmanrocks/Jacobian	David Freifeld	1	+75	-54
2021-03-29 18:37	Add python usage example	David Freifeld	1	+38	-0
2021-03-29 18:32	Add submodules to python bindings	David Freifeld	1	+75	-54
2021-03-29 03:39	Edit examples in README	richardfeynmanrocks	1	+17	-35
2021-03-29 03:28	Network interactivity aids and visualization in demo	David Freifeld	3	+36	-3
2021-03-29 02:23	Port decay functions and respective python bindings	David Freifeld	3	+56	-5
2021-03-29 02:03	C++ and formatting tweaks	David Freifeld	6	+23	-24
2021-03-29 01:36	Update Python bindings to use optimizers	David Freifeld	1	+41	-33
2021-03-29 01:17	Removed activation name property, fixed optimizers	David Freifeld	4	+77	-87
2021-03-28 06:46	Futher progress with Python binds	David Freifeld	2	+38	-28
2021-03-28 05:58	Merge branch 'master' of https://github.com/richardfeynmanrocks/Jacobian	David Freifeld	1	+3	-18
2021-03-28 05:58	Successful build of Python bindings	David Freifeld	3	+30	-25
2021-01-24 23:53	README edits	richardfeynmanrocks	1	+3	-18
2021-01-18 22:19	Minor debugging edits	David Freifeld	5	+19	-46
2021-01-18 22:17	Rational function approximation of tanh	David Freifeld	1	+10	-11
2021-01-18 21:32	Initial optimizations of activation functions	David Freifeld	2	+33	-13
2020-12-29 03:28	Trimming unnecessary files in repo	David Freifeld	10	+0	-550
2020-12-29 03:17	Cleanup to silence warnings, edited Eigen includes for Arch	David Freifeld	7	+35	-106
2020-11-29 21:07	Added UNLICENSE	richardfeynmanrocks	1	+24	-0
2020-11-29 19:09	Removed traces of AVX calls from master	David Freifeld	6	+2	-108
2020-11-29 18:56	Removed traces of PReLU from master	David Freifeld	3	+2	-25
2020-11-29 18:38	Modularized backpropagation to a degree	David Freifeld	3	+10	-51
2020-11-29 18:22	Moved cwise_product() and other macros to bpnn.hpp	David Freifeld	2	+14	-14
2020-11-29 18:21	Continued struggles with exploding gradients	David Freifeld	2	+156	-146
2020-11-29 17:36	Fixed basic convolutional backprop	David Freifeld	1	+3	-4
2020-11-29 17:06	Swapped convolution operation for CNN backprop	David Freifeld	1	+11	-9
2020-11-29 16:52	Attempted rewrite of CNN backpropagation	David Freifeld	1	+24	-46
2020-11-25 07:05	Indendation + debugging	David Freifeld	1	+88	-88
2020-11-09 05:53	Merge branch 'master' of https://github.com/richardfeynmanrocks/Jacobian	David Freifeld	1	+2	-0
2020-11-09 05:52	Fixed CNN errors until it compiled	David Freifeld	3	+7	-3
2020-11-07 10:15	Attempted fix of compression errors	David Freifeld	1	+12	-10
2020-11-04 21:32	README linebreaks	richardfeynmanrocks	1	+1	-1
2020-11-04 21:30	More README tweaks	richardfeynmanrocks	1	+1	-1
2020-11-04 21:29	Another attempt to fix README	richardfeynmanrocks	1	+1	-5
2020-11-04 21:26	Maybe fixed README?	richardfeynmanrocks	1	+3	-6
2020-11-04 21:21	Badges?	richardfeynmanrocks	1	+9	-0
2020-10-12 03:23	Tweaks	David Freifeld	2	+22	-38
2020-10-12 01:55	Fix prep() and other tweaks	David Freifeld	2	+11	-15
2020-10-11 19:46	Compression better off experimental for now	David Freifeld	3	+7	-34
2020-10-11 19:15	Segfault on function call	David Freifeld	4	+43	-11
2020-10-11 18:53	Tentatively declaring file reading bugs fixed	David Freifeld	3	+17	-3
2020-10-11 18:38	Fixed some of program-breaking bug	David Freifeld	2	+2	-9
2020-10-10 19:16	Investigating error	David Freifeld	1	+16	-14
2020-10-10 17:55	Ludicrous file read speed achieved	David Freifeld	1	+52	-19
2020-09-28 19:11	Tweaks to buffer	David Freifeld	2	+7	-8
2020-09-28 01:43	File read is broken for no apparent reason	David Freifeld	2	+30	-16
2020-09-27 08:00	Banished C casts	David Freifeld	2	+5	-3
2020-09-27 07:31	Bugs persist, tweaks to read function	David Freifeld	3	+9	-13
2020-09-27 07:04	Tweaks + terrible bugs	David Freifeld	3	+22	-8
2020-09-27 02:42	Added exceptions and more	David Freifeld	2	+4	-5
2020-09-27 02:10	GSL Expects/Ensures because apparently C++ contracts don't exist yet	David Freifeld	3	+18	-5
2020-09-26 03:07	Update README	David Freifeld	1	+0	-15
2020-09-26 02:46	Everything is broken now	David Freifeld	5	+27	-97
2020-09-25 23:59	Small tweaks	David Freifeld	2	+5	-1
2020-09-25 23:59	Incorporating faster file read into BPNN	David Freifeld	3	+99	-10
2020-09-25 21:57	readv() is ultimately disappointing	David Freifeld	1	+9	-21
2020-09-25 21:38	Attribution + readv() improvements	David Freifeld	2	+20	-10
2020-09-25 21:16	Added readv() bench, problems with buffers	David Freifeld	1	+31	-8
2020-09-25 17:51	Added beginning of capsule networks	David Freifeld	1	+23	-0
2020-09-25 17:33	Created CNN header	David Freifeld	2	+50	-50
2020-09-24 04:38	Work with mmap()	David Freifeld	1	+24	-31
2020-09-23 17:00	Disable workflows for now	David Freifeld	2	+0	-72
2020-09-23 16:58	Testing use of fcntl with RDADVISE	David Freifeld	1	+42	-11
2020-09-22 22:02	Preparing for benchmarks	David Freifeld	1	+6	-11
2020-09-22 20:24	backing up	David Freifeld	1	+15	-11
2020-09-22 16:44	Bizarre shifts in file	David Freifeld	1	+9	-8
2020-09-22 06:33	Broken switch to avoiding newlines	David Freifeld	1	+10	-7
2020-09-22 05:29	Fixed benchmark	David Freifeld	1	+7	-5
2020-09-22 04:08	Binary gives lots of speed	David Freifeld	1	+7	-13
2020-09-22 03:35	Troubles with fwrite()	David Freifeld	1	+17	-12
2020-09-21 20:56	Attempt to fix workflow	richardfeynmanrocks	1	+1	-1
2020-09-22 03:12	Decided file read is not fast enough	David Freifeld	1	+74	-27
2020-09-21 18:40	Switched to faster file read, 30% overall speedup	David Freifeld	5	+23	-37
2020-09-21 17:40	Migrating network to 3x faster file read	David Freifeld	5	+51	-24
2020-09-21 06:03	Things work now	David Freifeld	1	+4	-7
2020-09-21 05:37	Updated scan function	David Freifeld	1	+25	-8
2020-09-21 04:21	Fileread continues to be a not straightforward microoptimization that will ul...	David Freifeld	1	+16	-18
2020-09-21 03:32	CNN tweaks and more fileread business	David Freifeld	2	+18	-22
2020-09-21 01:33	More difficulties with adding faster file read	David Freifeld	1	+26	-8
2020-09-20 20:51	Headway into faster file read + parse	David Freifeld	1	+25	-3
2020-09-20 19:39	Check fixes + more experiments with file read	David Freifeld	2	+42	-51
2020-09-07 23:02	Fixed basic RNN runtime errors	David Freifeld	1	+2	-11
2020-09-07 22:54	Beginning to fix RNN issues	David Freifeld	1	+34	-7
2020-09-07 05:26	Fixed all RNN compile-time errors, now for runtime errors	David Freifeld	3	+22	-8
2020-09-07 04:54	Split softmax into its own function + more RNN setup	David Freifeld	3	+54	-23
2020-09-07 04:13	Cleaning up ANN + more setup for RNN	David Freifeld	3	+19	-9
2020-09-07 03:56	Beginning RNN implementation	David Freifeld	1	+43	-0
2020-09-05 18:56	Fixes to layer add function	David Freifeld	5	+31	-29
2020-09-05 18:25	Minimal updates, better typecasting	David Freifeld	8	+142	-212
2020-08-12 00:04	Various tweaks	David Freifeld	4	+9	-12
2020-08-11 22:57	Added basic checks + custom checks added to workflow	David Freifeld	6	+130	-51
2020-08-11 22:04	Basic ability to call checks	David Freifeld	5	+188	-175
2020-08-11 21:42	Isolating individual checks to retain some functionality	David Freifeld	3	+103	-132
2020-08-11 20:23	Working on checks some more	David Freifeld	4	+172	-179
2020-08-11 02:23	Continued frustrations with copy constructor	David Freifeld	4	+36	-8
2020-08-11 01:45	Working on proper copy constructors to allow for safer checks	David Freifeld	4	+194	-182
2020-08-11 01:04	Continued struggles with pip packaging	David Freifeld	1	+5	-6
2020-08-09 23:06	Wow setuptools can be annoying	David Freifeld	2	+50	-26
2020-08-09 22:38	Working on python packaging	David Freifeld	5	+30	-15
2020-08-08 02:20	More stability in checks + code	David Freifeld	6	+21	-56
2020-08-08 01:23	Overall improvements to quality of CMake build	David Freifeld	6	+56	-324
2020-08-07 21:45	Slowly moving towards smoother python build + install	David Freifeld	3	+23	-24
2020-08-07 21:22	Working on proper python install process	David Freifeld	4	+15	-17
2020-08-07 20:43	Added new checks + fixed CMakeLists	David Freifeld	3	+34	-3
2020-08-07 20:26	Began to replicate original Makefile	David Freifeld	2	+25	-14
2020-08-07 20:03	Yet another edit to workflow file	richardfeynmanrocks	1	+4	-6
2020-08-07 19:58	Removed install.sh, ignoring pybind install for now	David Freifeld	1	+15	-3
2020-08-07 19:49	More setup work	David Freifeld	4	+18	-5
2020-08-07 19:33	Github Action experiments	richardfeynmanrocks	1	+21	-0
2020-08-07 19:32	Attempt to use CMake	David Freifeld	3	+281	-34
2020-08-07 18:43	Restructuring directory	David Freifeld	9	+0	-0
2020-08-07 06:18	Addressed change of plans + tweaks	David Freifeld	4	+13	-4
2020-08-07 04:42	Added ability to toggle explicit AVX functions	David Freifeld	3	+23	-14
2020-08-07 03:33	A little bit of iteration fixes bugs	David Freifeld	1	+7	-0
2020-08-07 02:53	Broken AVX functions are broken	David Freifeld	2	+20	-38
2020-08-07 02:36	Actual code changes	David Freifeld	1	+3	-23
2020-08-07 02:35	Speed improvements to AVX product	David Freifeld	1	+1	-4
2020-08-06 20:55	Benchmarks show copy is problematic	David Freifeld	2	+19	-7
2020-08-06 20:22	Tweaking to make things work	David Freifeld	6	+30	-16
2020-08-06 07:26	More repetitive AVX functions	David Freifeld	1	+63	-0
2020-08-06 07:18	Fixes + AVX exp	David Freifeld	3	+15	-26
2020-08-06 06:35	Removed memory leak + debugging	David Freifeld	2	+22	-9
2020-08-06 01:27	Mystery memory leak isnt in AVX	David Freifeld	1	+8	-4
2020-08-06 01:00	Update to reflect future plans	David Freifeld	2	+24	-35
2020-08-06 00:20	Copious amounts of errors	David Freifeld	3	+12	-10
2020-08-05 23:25	Began custom AVX functions for speedups	David Freifeld	4	+515	-497
2020-08-04 01:26	Added early stopping + fixed makefile	David Freifeld	5	+19	-11
2020-08-04 01:03	Adding progress bar	David Freifeld	2	+28	-3
2020-08-04 00:10	Add nesterov momentum	David Freifeld	4	+20	-3
2020-08-03 20:13	Messing with OpenMP, MKL, BLAS and LAPACK	David Freifeld	3	+11	-20
2020-08-03 18:56	Attempts to use old mapreduce code	David Freifeld	4	+87	-126
2020-08-02 18:38	Attempts to repair CNN backprop	David Freifeld	3	+48	-31
2020-08-02 17:27	Tweak to naming convention	David Freifeld	3	+15	-15
2020-08-02 02:31	Update decays (they're broken now)	David Freifeld	4	+29	-15
2020-08-02 02:00	Fixed demon + restructuring	David Freifeld	3	+64	-52
2020-08-02 01:32	Adamax goes boom	David Freifeld	2	+6	-5
2020-08-02 01:08	Adam no longer bugs out	David Freifeld	2	+2	-2
2020-08-02 00:06	Numerical gradients seem to be highly incorrect	David Freifeld	3	+18	-10
2020-08-01 23:32	Working on gradient check	David Freifeld	3	+45	-3
2020-08-01 19:55	Added demon + more NaNs	David Freifeld	2	+20	-8
2020-08-01 19:15	Added AdaMax + found adam bugs	David Freifeld	1	+16	-3
2020-08-01 18:47	Added Adam prototype	David Freifeld	3	+23	-6
2020-08-01 17:55	Added method of setting optimizer	David Freifeld	3	+29	-8
2020-08-01 02:08	Quick readme update	David Freifeld	1	+2	-2
2020-08-01 02:05	Add more sklearn benchmarks	David Freifeld	5	+14	-5
2020-07-31 23:14	Testing momentum - weight updates exploding...	David Freifeld	2	+15	-6
2020-07-31 20:49	Added new figure + readme tweak	David Freifeld	2	+2	-2
2020-07-31 20:39	Benchmark code update to include more metrics	David Freifeld	4	+44	-19
2020-07-31 06:34	Benchmark update + better figures	David Freifeld	10	+29	-12
2020-07-31 06:31	Update Makefile to remove unecessary flags	David Freifeld	1	+2	-2
2020-07-31 04:49	Tweaks + readme update	David Freifeld	3	+25	-28
2020-07-31 00:38	Pursuing eigen functions in softmax	David Freifeld	1	+38	-0
2020-07-31 00:15	More investigating cwise functions	David Freifeld	2	+36	-9
2020-07-30 23:26	Experiments show cwise functions are faster	David Freifeld	1	+21	-0
2020-07-30 22:59	L1 regularization added - may be sketch	David Freifeld	4	+24	-11
2020-07-30 22:17	PReLU now supported (Python is buggy tho)	David Freifeld	3	+5	-7
2020-07-30 21:50	Located PReLU bug	David Freifeld	2	+13	-6
2020-07-30 02:53	PReLU very very buggy	David Freifeld	2	+11	-9
2020-07-30 02:19	PReLU now compiles	David Freifeld	3	+25	-17
2020-07-30 02:00	More progress on PReLU	David Freifeld	1	+10	-24
2020-07-30 01:27	Beginnings of PReLU	David Freifeld	2	+22	-3
2020-07-30 01:01	New activations are now allowed to be added	David Freifeld	1	+17	-0
2020-07-30 00:28	More activations	David Freifeld	2	+29	-0
2020-07-27 22:27	Attempt to show new bench fig	David Freifeld	2	+10	-10
2020-07-27 22:23	Update benchmark	David Freifeld	5	+18	-29
2020-07-27 22:08	More tweaks	David Freifeld	5	+112	-101
2020-07-27 21:38	Updated scikit demo	David Freifeld	1	+29	-13
2020-07-27 21:20	Working on more benchmarks	David Freifeld	3	+122	-0
2020-07-26 03:36	Various tweaks + moving to tensors?	David Freifeld	6	+40	-45
2020-07-15 23:40	CNN pooling is broken	David Freifeld	2	+22	-19
2020-07-15 19:49	Large bugfix + much better accuracy	David Freifeld	2	+10	-16
2020-07-15 19:31	Backprop through more than one conv layer works	David Freifeld	1	+32	-22
2020-07-15 03:33	Attempt at using matrix functions reveals they are slow	David Freifeld	2	+35	-0
2020-07-15 02:42	Working on speedups + icpc works without MKL	David Freifeld	5	+40	-38
2020-07-14 23:48	CNN doesn't break if you don't breathe on it too hard	David Freifeld	4	+60	-18
2020-07-14 23:01	Backprop looks normal for BPNN	David Freifeld	5	+21	-26
2020-07-14 19:00	Added notice to code	David Freifeld	16	+128	-4
2020-07-14 18:44	Makefile update	David Freifeld	1	+3	-3
2020-07-14 16:04	Backprop is unstable + largely broken	David Freifeld	4	+50	-33
2020-07-13 18:57	CNN has been possessed by NaNs	David Freifeld	1	+115	-59
2020-07-13 17:39	Working on more safety checks	David Freifeld	3	+27	-23
2020-07-13 16:53	Working on safety with errors	David Freifeld	3	+31	-5
2020-07-13 15:31	Added to README	richardfeynmanrocks	1	+60	-17
2020-07-12 22:58	Further scrutiny reveals big backprop issues	David Freifeld	3	+24	-10
2020-07-12 19:22	Training no longer errors/nans but is very broken	David Freifeld	2	+34	-25
2020-07-12 18:47	Linear output seems to annihilate softmax (and network)	David Freifeld	2	+17	-12
2020-07-12 18:20	Backprop doesn't segfault	David Freifeld	2	+16	-25
2020-07-12 17:54	Loss+accuracy tentatively work now	David Freifeld	2	+15	-7
2020-07-12 17:36	Added crossentropy cost function	David Freifeld	1	+12	-6
2020-07-12 06:15	Switching to softmax	David Freifeld	2	+14	-4
2020-07-12 00:34	SVM loss causes exploding gradients	David Freifeld	2	+23	-6
2020-07-12 00:08	Working on backprop fixes	David Freifeld	2	+21	-9
2020-07-11 22:37	Attempting switch to SVM loss	David Freifeld	2	+20	-9
2020-07-11 19:16	Using MNIST data	David Freifeld	1	+55	-15
2020-07-11 18:47	Attempts to fix checks result in more bugs	David Freifeld	4	+14	-15
2020-07-11 18:39	CNN with custom kernel now compiles	David Freifeld	1	+41	-41
2020-07-11 17:47	Added decay function support	David Freifeld	2	+37	-11
2020-07-11 17:23	Experiments with regularization + momentum	David Freifeld	2	+9	-3
2020-07-11 05:45	Got sweeps working	David Freifeld	2	+18	-7
2020-07-11 03:09	Reorganized directory.	David Freifeld	22	+603	-606
2020-07-11 02:39	Separated checks from main code	David Freifeld	3	+163	-145
2020-07-11 02:25	2 more sanity checks	David Freifeld	1	+42	-3
2020-07-11 02:03	Added floating point sanity check	David Freifeld	1	+17	-2
2020-07-11 01:44	More sanity checks	David Freifeld	2	+45	-5
2020-07-11 01:03	Working on regularization + sanity checks	David Freifeld	3	+48	-25
2020-07-11 00:13	Extending float changes to strassen + utils	David Freifeld	2	+38	-37
2020-07-11 00:01	Working on export as actual python lib	David Freifeld	2	+22	-0
2020-07-10 23:28	Attempts to update benchmark are met with segfualts	David Freifeld	1	+7	-5
2020-07-10 23:16	Attempts to use wandb sweep	David Freifeld	2	+56	-27
2020-07-10 22:51	Setup example.py to use regularization	David Freifeld	3	+10	-8
2020-07-10 22:31	Broken and bad numerical gradient	David Freifeld	2	+37	-14
2020-07-10 22:09	Working L2 regularization in backprop?	David Freifeld	3	+8	-7
2020-07-10 22:02	Switch to floats for speedups	David Freifeld	4	+36	-52
2020-07-10 21:38	Beginnings of gradient check	David Freifeld	3	+38	-58
2020-07-07 23:57	Further attempts to locate bug	David Freifeld	1	+5	-5
2020-07-07 02:11	Struggles to resolve MR bug	David Freifeld	2	+5	-5
2020-07-05 18:14	CNN now stabler	David Freifeld	1	+23	-15
2020-07-05 17:57	Some amount of training present	David Freifeld	1	+6	-3
2020-07-05 17:49	Problematic backprop	David Freifeld	2	+20	-18
2020-07-05 17:49	Remove extra function	David Freifeld	1	+17	-35
2020-07-04 23:49	Added nicer list_net() to BPNN	David Freifeld	1	+4	-3
2020-07-04 22:48	Continued problems w/ strassen	David Freifeld	1	+3	-1
2020-07-04 17:37	Attempt to get strassen pad working	David Freifeld	2	+25	-5
2020-07-04 05:58	Updated scripts to try and work w/ strassen	David Freifeld	3	+37	-34
2020-07-04 05:57	Attempt to use strassen for feedforward	David Freifeld	3	+30	-5
2020-07-04 05:33	Reorganizing tests.	David Freifeld	3	+59	-60
2020-07-04 05:26	Strassen's algorithm is functional	David Freifeld	1	+7	-14
2020-07-04 02:58	Padding is functional	David Freifeld	1	+11	-5
2020-07-04 01:11	Padding causes segfaults...	David Freifeld	1	+9	-4
2020-07-04 00:46	Added bias to conv layer + bad training of bias	David Freifeld	1	+6	-3
2020-07-04 00:29	Sort of trains sometimes	David Freifeld	1	+13	-7
2020-07-03 23:42	Kernel update in right shape, still horribly broken	David Freifeld	1	+6	-3
2020-07-03 23:08	Backprop is very broken	David Freifeld	1	+13	-5
2020-07-03 22:46	Beginnings of backprop	David Freifeld	1	+29	-8
2020-07-03 20:15	Slight fixes + more in list_net()	David Freifeld	3	+24	-24
2020-07-03 19:57	More feedforward fixed	David Freifeld	1	+11	-10
2020-07-03 19:28	Broken convolution? Also CNN backprop confusing	David Freifeld	1	+12	-4
2020-07-03 18:27	list_net() now actually readable	David Freifeld	1	+5	-5
2020-07-03 18:04	Specific list_net() for CNN	David Freifeld	1	+21	-6
2020-07-03 02:19	Using library for bad network viz	David Freifeld	1	+38	-0
2020-07-03 01:04	Successful feedforward.	David Freifeld	1	+9	-11
2020-07-03 00:57	CNN executes	David Freifeld	1	+16	-6
2020-07-03 00:01	Closer to proper execution	David Freifeld	1	+11	-8
2020-07-02 23:35	CNN getting closer to running	David Freifeld	1	+27	-13
2020-07-02 23:05	CNN compiles	David Freifeld	1	+18	-14
2020-07-02 18:08	Checking if lambda is problem	David Freifeld	1	+36	-18
2020-07-02 03:16	Attempts to improve Makefile	David Freifeld	1	+11	-3
2020-07-02 03:09	Strassen's algorithm continues to drain my soul	David Freifeld	1	+7	-4
2020-07-02 02:34	Update to W&B demo	David Freifeld	2	+17	-9
2020-07-02 02:14	Validation step now functional	David Freifeld	4	+8	-12
2020-07-02 01:51	Struggling to defeat evil carriage return	David Freifeld	1	+8	-4
2020-07-02 01:11	Semi-functional test function, working split	David Freifeld	2	+59	-15
2020-07-02 00:35	Beginnings of nicer testing	David Freifeld	2	+26	-32
2020-07-01 23:17	Some W&B tweaks	David Freifeld	4	+29	-14
2020-07-01 21:55	Using wandb to measure things more properly	David Freifeld	5	+32	-63
2020-07-01 19:55	Attempted fixes to no avail	David Freifeld	1	+9	-3
2020-07-01 19:20	Strassen now works but has incorrect quadrants	David Freifeld	1	+7	-14
2020-07-01 19:03	Compiles, but both are so slow test is not done	David Freifeld	1	+19	-9
2020-07-01 18:44	Some fixes to errors, one remains	David Freifeld	1	+28	-23
2020-07-01 18:26	Broken strassen prototype	David Freifeld	1	+19	-2
2020-07-01 04:27	Markdown bulleted lists are bad	richardfeynmanrocks	1	+5	-5
2020-07-01 04:23	Better README, Makefile, and a bit more Strassen	David Freifeld	3	+43	-15
2020-07-01 04:00	Beginnings of Strassen tests	David Freifeld	1	+25	-0
2020-07-01 03:59	A little more benchmarking + Eigen fix	David Freifeld	2	+23	-24
2020-07-01 03:26	Boilerplate preprocess + avoid extra mem allocation	David Freifeld	1	+18	-14
2020-07-01 03:11	Added constructors + funcs for adding layers	David Freifeld	1	+56	-2
2020-07-01 02:55	Some fixes, realized CNNs will be formidable	David Freifeld	1	+10	-6
2020-07-01 02:16	Working more w/ CNNs	David Freifeld	2	+46	-2
2020-07-01 00:15	Loose boilerplate for CNN	David Freifeld	1	+30	-0
2020-06-30 19:40	Thrilling development reveals I can't read	David Freifeld	1	+1	-1
2020-06-30 19:23	Further compiler optimizations	David Freifeld	1	+1	-1
2020-06-30 18:04	Significant speedup	David Freifeld	3	+38	-37
2020-06-30 06:08	Compiling has exploded, several bugs noticed	David Freifeld	2	+5	-18
2020-06-30 03:45	Added nice-looking banner?	David Freifeld	6	+5	-5
2020-06-29 18:30	Brief clarification to README	richardfeynmanrocks	1	+1	-1
2020-06-29 18:27	Added pictures	David Freifeld	3	+0	-0
2020-06-29 18:26	Updated README	David Freifeld	3	+45	-35
2020-06-29 16:35	Poor attempts to benchmark more	David Freifeld	4	+50	-53
2020-06-29 06:09	Ran benchmarking tests	David Freifeld	3	+74	-33
2020-06-29 02:35	Screw it, no Adagrad	David Freifeld	3	+25	-23
2020-06-29 00:43	Adagrad struggles	David Freifeld	4	+20	-9
2020-06-28 23:39	Experiments with Adagrad	David Freifeld	3	+8	-5
2020-06-28 22:13	Switch to icpc and more compile flags	David Freifeld	3	+7	-46
2020-06-28 18:39	More work on speed + benchmarking	David Freifeld	5	+24	-12
2020-06-28 16:40	Much faster string parse	David Freifeld	1	+8	-15
2020-06-28 06:04	Trying to speedup scanf	David Freifeld	3	+47	-34
2020-06-28 04:50	More elaborate build configurations	David Freifeld	1	+12	-7
2020-06-28 03:29	Working on compiler-side speedups	David Freifeld	2	+17	-4
2020-06-28 02:05	Oops turns out ifstream is slower	David Freifeld	3	+26	-21
2020-06-28 01:35	Working on speeding up file i/o	David Freifeld	4	+17	-18
2020-06-27 21:56	Various experiments	David Freifeld	6	+41	-30
2020-06-27 03:08	Found activation bugs, fixed some + added new activation	David Freifeld	3	+9	-5
2020-06-27 02:41	Experiments with SGD w/ momentum	David Freifeld	3	+11	-10
2020-06-27 01:05	Buggy mapreduce call	David Freifeld	1	+20	-13
2020-06-27 00:37	Incorporating MR update + 3 new activations	David Freifeld	4	+29	-4
2020-06-26 22:53	Backup for posterity	David Freifeld	1	+7	-7
2020-06-26 20:05	Untested demo for NetworkArray	David Freifeld	1	+74	-3
2020-06-26 19:43	Adding more general parallel network array	David Freifeld	2	+16	-7
2020-06-26 19:17	More fixes, working on speedups	David Freifeld	7	+30	-32
2020-06-26 03:51	Added plenty of activation functions	David Freifeld	5	+49	-35
2020-06-26 03:05	Benchmarking in hopes of speedup	David Freifeld	5	+38	-8
2020-06-26 02:22	Nicely working python lib	David Freifeld	6	+158	-140
2020-06-26 01:35	Rectifier works, custom activations essentially allowed	David Freifeld	5	+35	-37
2020-06-26 01:08	Trying to add functional-style rectifier	David Freifeld	3	+24	-44
2020-06-25 21:26	Added Xavier initialization?	David Freifeld	2	+10	-5
2020-06-25 20:03	New construction works properly	David Freifeld	2	+10	-21
2020-06-25 19:45	Full boilerplate, lots of segfaults	David Freifeld	2	+17	-66
2020-06-25 19:38	Beginnings of nicer network construction	David Freifeld	2	+19	-30
2020-06-25 19:16	Successful training of biases	David Freifeld	2	+5	-4
2020-06-25 19:05	All layers now updated: training has decaying accuracy	David Freifeld	2	+10	-11
2020-06-25 18:30	Further tweaks to fix backprop	David Freifeld	2	+7	-5
2020-06-25 17:57	Basic backprop fixed	David Freifeld	2	+10	-10
2020-06-25 06:25	Attempts at backprop fixes	David Freifeld	2	+21	-8
2020-06-25 04:34	Restructuring for biases	David Freifeld	1	+8	-7
2020-06-25 02:21	Forgot to add correct files...	David Freifeld	2	+29	-0
2020-06-25 02:19	Added proper activation setting function	David Freifeld	2	+25	-1
2020-06-25 01:57	Bug tentatively marked as fixed	David Freifeld	1	+7	-4
2020-06-25 01:43	Custom activation functions but horribly buggy	David Freifeld	3	+49	-9
2020-06-25 01:08	Still working on switch	David Freifeld	3	+11	-8
2020-06-24 23:53	Working on custom activation functions	David Freifeld	3	+27	-37
2020-06-24 18:40	Switched back to FILE*	David Freifeld	2	+11	-20
2020-06-23 19:25	Switching to using istream	David Freifeld	4	+46	-11
2020-06-23 18:55	Experiments with mmap	David Freifeld	1	+39	-0
2020-06-23 18:24	Continued attempts to fix speed	David Freifeld	1	+11	-5
2020-06-23 02:03	Broke training but sped up batch advancement	David Freifeld	3	+16	-21
2020-06-23 00:59	Merge python-bindings	David Freifeld	5	+75	-12
2020-06-23 00:46	Usability + minimal Makefile	David Freifeld	4	+48	-51
2020-06-23 00:06	Making mr_bpnn_2 more usable	David Freifeld	1	+54	-14
2020-06-22 23:24	First successful test of python bindings.	David Freifeld	2	+13	-2
2020-06-22 21:22	Remove old c code	David Freifeld	1	+0	-16
2020-06-22 21:21	Fixes to readme and execution of demos	David Freifeld	3	+5	-9
2020-06-22 21:18	Added readme	David Freifeld	1	+34	-0
2020-06-22 17:52	Beginnings of pybind11	David Freifeld	1	+17	-3
2020-06-22 05:22	Some cleaning up	David Freifeld	1	+15	-39
2020-06-22 03:09	Fixed training, no clue why it helped	David Freifeld	1	+1	-1
2020-06-22 01:34	Further attempts to improve training	David Freifeld	2	+17	-5
2020-06-21 22:40	Quick fix	David Freifeld	1	+7	-5
2020-06-21 20:02	Trains in parallel (with bugs)	David Freifeld	3	+34	-38
2020-06-21 18:52	Rectified sigmoid seems to work nicer	David Freifeld	2	+15	-9
2020-06-21 18:21	Continuing work on fixing net	David Freifeld	3	+16	-10
2020-06-21 02:09	Training batch advancement works properly	David Freifeld	1	+18	-19
2020-06-21 01:51	Batch advancement almost works	David Freifeld	1	+12	-18
2020-06-21 01:16	Trying to fix file reading	David Freifeld	3	+39	-33
2020-06-20 17:57	Keras demo comment and fix	David Freifeld	1	+12	-1
2020-06-20 00:43	More stats outputs, determined rough source of a bug	David Freifeld	2	+18	-8
2020-06-20 00:06	MR_BPNN_2 sort of works, lots and lots of bugs	David Freifeld	1	+17	-16
2020-06-19 21:54	Broken prototype of MR_BPNN_2	David Freifeld	1	+107	-0
2020-06-19 21:38	Manually tracking how long parts of epoch take	David Freifeld	1	+19	-5
2020-06-19 20:39	Fixed some of batch advancement problems	David Freifeld	2	+30	-6
2020-06-19 20:08	Issues with loss/accuracy calculations	David Freifeld	3	+30	-39
2020-06-19 19:37	More benchmarking for comparison	David Freifeld	3	+31	-11
2020-06-19 19:06	Added keras benchmark code	David Freifeld	1	+26	-0
2020-06-19 17:28	Working towards fixing mr_bpnn_1 and writing mr_bpnn_2	David Freifeld	3	+23	-21
2020-06-19 01:10	Working mr_bpnn_1 if you ignore the memory leaks	David Freifeld	1	+28	-7
2020-06-18 18:02	Fixed dereferencing bug	David Freifeld	1	+9	-6
2020-06-18 03:54	Attempts to get mr_bpnn_1 working	David Freifeld	2	+19	-19
2020-06-17 01:44	Also fixed header	David Freifeld	1	+56	-0
2020-06-17 01:43	Restructuring, strange malloc bug	David Freifeld	5	+359	-547
2020-06-16 20:03	Broken reduce function but parallel training	David Freifeld	1	+26	-12
2020-06-16 19:39	Working on messy MR proof of concept	David Freifeld	1	+82	-10
2020-06-16 02:15	Lack of sigmoid results in NaNs...	David Freifeld	1	+16	-11
2020-06-16 01:35	New simpler backprop behaves strangely	David Freifeld	1	+28	-34
2020-06-16 00:56	Fixing backprop and trying mapreduce	David Freifeld	2	+58	-86
2020-06-15 21:40	Sort of legitimately trains	David Freifeld	1	+7	-7
2020-06-15 19:52	Closer to working backprop	David Freifeld	1	+11	-12
2020-06-15 18:54	Fixing backprop: problems remain	David Freifeld	1	+41	-24
2020-06-15 01:47	Added test function	David Freifeld	1	+37	-3
2020-06-15 00:20	Backprop is horribly broken	David Freifeld	1	+9	-7
2020-06-14 23:55	Training is broken?	David Freifeld	1	+23	-21
2020-06-14 23:34	Working towards proper SGD	David Freifeld	1	+66	-33
2020-06-14 19:11	Broken backprop for all weights	David Freifeld	1	+24	-11
2020-06-14 05:35	Expanded training, ignoring blatant backprop flaws	David Freifeld	1	+51	-15
2020-06-14 04:51	Training...sort of.	David Freifeld	1	+37	-31
2020-06-14 00:12	Backprop is frustrating	David Freifeld	1	+37	-18
2020-06-13 22:41	Backprop is scary	David Freifeld	1	+11	-5
2020-06-13 05:14	Inching towards backprop	David Freifeld	1	+21	-5
2020-06-13 01:40	Building up to backprop	David Freifeld	1	+22	-3
2020-06-12 19:42	Fixed biases init, debug function	David Freifeld	1	+12	-9
2020-06-12 17:26	Fixed feedforward	David Freifeld	1	+4	-8
2020-06-12 16:44	Biases added.	David Freifeld	1	+14	-0
2020-06-12 05:55	Full init, broken feedforward	David Freifeld	1	+51	-14
2020-06-12 03:54	Successful parsing of input to matrix	David Freifeld	1	+19	-8
2020-06-12 02:29	Reworked to include more matrices	David Freifeld	1	+29	-16
2020-06-11 22:27	Fixed some issues, runs now	David Freifeld	1	+31	-29
2020-06-11 21:38	C++ implmentation begins	David Freifeld	1	+76	-0
2020-06-09 16:35	Fixed some problems, still momre	David Freifeld	1	+36	-27
2020-06-08 04:28	Plenty of malloc problems	David Freifeld	1	+46	-15
2020-06-07 22:57	Basic neural net initialization	David Freifeld	1	+56	-6
2020-06-07 18:00	Boilerplate / init commit	David Freifeld	1	+55	-0

	jacobian a basic keras-like neural network library for c++/python
	Log \| Files \| Refs \| README