rubygrad 1.2.2 → 1.2.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f05df7e74616deb2db9a7c7340e105e0997268d9d2938d07dbabf51395684fe5
-  data.tar.gz: 6179a5c6a57c37411784195a9ca3a15f73debdd06aaeae46a754ff5857e60770
+  metadata.gz: d02ad8e3dfa9fb89c29c21d12f17488ca1ddb8dbf2cbfdfdd61f110a84cc7b02
+  data.tar.gz: 2cb1da6e863aef973b12c69162723021c6bf6f07d9609628a54e889fdca1e362
 SHA512:
-  metadata.gz: 2187327cf7c8b2a8e323b90138c49b7bd6cb5c84ed0e784a5c288556525cf0f8c619dd2aef024b6afaa27aad5b4f2a104457e67dd7158a121f41e9b5dfa631c3
-  data.tar.gz: ebb19b5217fb756fa9072dcf9e2b39ff503f93ac94f782cc45e2ffdc73ee85e6f5c9e471d7995d1017aa0697b97437cf5eb56be1cee34d79e30b0aaf4d5818b9
+  metadata.gz: 0bf4727a51a73b307686f219ce5633880c9ee80348ca1834ae9f4b357a5d20bef9c572f7ac6b999a4053b9cc2229336de5ddcc9e2e6aa475bb246bbad5dae4fa
+  data.tar.gz: 0d08776331371d9cc65b87d099443af44ae8be5cb8b85f0596a680ec40205bf3514cf2a15e45e4433004a2a6a87b0341a7b52fb61d48cef3c491bad647833071

data/lib/nn.rb

@@ -50,6 +50,8 @@ class Neuron
             sum.relu
         elsif @activation_function == :sigmoid
             sum.sigmoid
+        elsif @activation_function == :none
+            sum
         else
             raise "Unsupported activation function: #{activation_function}"
         end
@@ -210,4 +212,11 @@ class MLP
         end
         out.size == 1 ? out[0] : out # for convenience
     end
+
+    def print_pass(learning_rate, loss, pass, passes, learning_rate_precision = 2, loss_precision = 10)
+        passes_format = "%#{passes.digits.length}d"
+        learning_rate_format = "%.#{learning_rate_precision}f"
+        loss_format = "%.#{loss_precision}f"
+        puts "Pass #{passes_format % (pass + 1)} => Learning rate: #{learning_rate_format % learning_rate} => Loss: #{loss_format % loss.value}"
+    end
 end

data/mlp_example.rb

@@ -27,10 +27,6 @@ y_expected = [1.0, -1.0, -1.0, 1.0]
 passes = 2000
 learning_rate = 0.2
 
-_loss_precision = 10
-_passes_format = "%#{passes.digits.length}d"
-_loss_format = "%.#{_loss_precision}f"
-
 (0...passes).each do |pass| 
 
     # forward pass (calculate output)
@@ -47,7 +43,8 @@ _loss_format = "%.#{_loss_precision}f"
     # improve neural net (update weights and biases)
     nn.parameters.each { |p| p.value -= learning_rate * p.grad }
 
-    puts "Pass #{_passes_format % (pass + 1)} => Learning rate: #{"%.10f" % learning_rate} => Loss: #{_loss_format % loss.value}" if (pass + 1) % 100 == 0 or pass == 0
+    # print some info about our progress from time to time
+    nn.print_pass(learning_rate, loss, pass, passes) if (pass + 1) % (passes / 20) == 0 or pass == 0
 
     break if loss.value == 0 # just for fun and just in case
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rubygrad
 version: !ruby/object:Gem::Version
-  version: 1.2.2
+  version: 1.2.4
 platform: ruby
 authors:
 - Sergio Oliveira Jr
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-03-22 00:00:00.000000000 Z
+date: 2023-03-23 00:00:00.000000000 Z
 dependencies: []
 description:
 email: sergio.oliveira.jr@gmail.com