crystal-fann/Lobby

Support and Discussion for crystal-fann shard

Bar Hofesh

@bararchy

Also, a quick question, which of the examples didn't work?

Martyn Jago

@mjago

Specs work

Finished in 71.3 milliseconds
14 examples, 0 failures, 0 errors, 0 pending

I compiled github/libfann head

Bar Hofesh

@bararchy

hm... running specs on master shows

Fann::Network
  initializes
works
Fann::Network
  initializes standard network
  initializes cascade network
  free memory
  trains on single data
  Shows MSE
  trains and evaluate single data
  train on batchMax epochs     8000. Desired error: 0.0010000000.
Epochs            1. Current error: 0.2745950818. Bit fail 2.
Epochs           91. Current error: 0.0009867755. Bit fail 0.
  train on batch
  train on cascadeMax neurons 500. Desired error: 0.001000
Neurons       0. Current error: 0.250000. Total error:  1.0000. Epochs    51. Bit fail   4
Neurons       1. Current error: 0.000746. Total error:  0.0030. Epochs   117. Bit fail   0. candidate steepness 1.00. function FANN_GAUSSIAN_SYMMETRIC
Train outputs    Current error: 0.000000. Epochs    123
  train on cascade

Finished in 3.2 milliseconds
10 examples, 0 failures, 0 errors, 0 pending

So only 10 examples

how did you get 14?

are you using latest version of Crystal-fann?

Martyn Jago

@mjago

:smile:

Ok so I was using a shard to get the error, and cloned your repo to run the specs

Bar Hofesh

@bararchy

Ok, so I'll just make sure again, are you using latest crystal-fann ?

maybe the shard isn't using latest crystal-fann version?

Martyn Jago

@mjago

NeuraLegion/crystal-fann

shard is using that with branch: master

I get this .......Max epochs 8000. Desired error: 0.0010000000.
Epochs 1. Current error: 0.3286948800. Bit fail 2.
Epochs 63. Current error: 0.0009598084. Bit fail 0.
.Max neurons 500. Desired error: 0.001000
Neurons 0. Current error: 0.250000. Total error: 1.0000. Epochs 51. Bit fail 4
Neurons 1. Current error: 0.000618. Total error: 0.0025. Epochs 116. Bit fail 0. candidate steepness 0.25. function FANN_GAUSSIAN_SYMMETRIC
Train outputs Current error: 0.000000. Epochs 120
......

Finished in 79.43 milliseconds
14 examples, 0 failures, 0 errors, 0 pending
Sun Dec 17: crystal-fann/ >
```

I hate gitter sometimes

Bar Hofesh

@bararchy

@mjago You're right, it was my bad, I wans't on latest

can you tell me which of the README examples wont work for you?

basiclly all of them are in the specs, so it's wierd they work there but not in the code

Martyn Jago

@mjago

# Work on array of test data (batch)

Hmm strange - one of the specs failed once (train_data_spec.cr I think) but now all pass - I’ll go and try the example again

Bar Hofesh

@bararchy

Sure, let me know :)

Martyn Jago

@mjago

Still fails yet the first and last example work fine

Perhaps its the crystal version ?

Crystal 0.24.1+5 [a1e90f0bd] (2017-12-16)

LLVM: 4.0.0
Default target: x86_64-apple-macosx

Bar Hofesh

@bararchy

I'm also on Head

I'm on Linux hough

should not matter, but let me do another test

Martyn Jago

@mjago

libFann version (head)

d71d547 * master origin/master ensure that cmake is not dependent on directory it's run in

Bar Hofesh

@bararchy

@mjago Looking at it more deeply, we didn't update the README to showcase latest specs

@mjago Updated the README now

can you try the changed examples and let me know if you get into any issue?

Martyn Jago

@mjago

@bararchy yay 🎉 all good now

Bar Hofesh

@bararchy

Cool, sorry for that XD

@mjago

enjoy the lib

Martyn Jago

@mjago

@barachy note however that without branch: master crystal deps installs an old version (with typos)

Bar Hofesh

@bararchy

I'll create a release

Martyn Jago

@mjago

@bararchy yes I know nothing about ANN etc so have some learning to do :+1:

@barachy thanks for use of the library !

Bar Hofesh

@bararchy

no problem, enjoy :)

Martyn Jago

@mjago

@bararchy I noticed that fann_set_learning_rate() expects to pass a float (doesn’t work with a double as currently defined)

Bar Hofesh

@bararchy

Hm..... interesting, let me take a look, I might missed something

Martyn Jago

@mjago

Do you want a PR for #learning_rate and #learning_rate = ?

Bar Hofesh

@bararchy

Sure, would love it :)

also, you should really take a look at SHAInet, it's much faster, and give better results, it also allows for more precision in network architecture

Martyn Jago

@mjago

Sure, I definitely will do

Bar Hofesh

@bararchy

:+1:

Martyn Jago

@mjago

My pet project is decoding morse from a noisy audio source - already works from a clean ‘ideal source'

Bar Hofesh

@bararchy

really cool! is that an opensource project?

Where communities thrive

People

Repo info

Activity