Add filters slider

chatbot_constructor.py

@@ -15,7 +15,7 @@ os.mkdir("cache")
 def hash_str(data: str):
     return hashlib.md5(data.encode('utf-8')).hexdigest()
 
-def train(message: str = "", epochs: int = 16, learning_rate: float = 0.001, emb_size: int = 128, inp_len: int = 16, data: str = ""):
+def train(message: str = "", epochs: int = 16, learning_rate: float = 0.001, emb_size: int = 128, inp_len: int = 16, kernels_count: int = 8, data: str = ""):
     data_hash = None
     if "→" not in data or "\n" not in data:
             if data in os.listdir("cache"):
@@ -39,7 +39,7 @@ def train(message: str = "", epochs: int = 16, learning_rate: float = 0.001, emb
         emb_layer = Embedding(input_dim=vocab_size, output_dim=emb_size, input_length=inp_len)(input_layer)
         attn_layer = MultiHeadAttention(num_heads=4, key_dim=128)(emb_layer, emb_layer, emb_layer)
         noise_layer = GaussianNoise(0.1)(attn_layer)
-        conv1_layer = Conv1D(64, 8, padding='same', activation='relu', strides=1, input_shape=(64, 128))(noise_layer)
+        conv1_layer = Conv1D(kernels_count, 8, padding='same', activation='relu', strides=1, input_shape=(64, 128))(noise_layer)
         conv2_layer = Conv1D(16, 4, padding='same', activation='relu', strides=1)(conv1_layer)
         conv3_layer = Conv1D(8, 2, padding='same', activation='relu', strides=1)(conv2_layer)
         flatten_layer = Flatten()(conv3_layer)
@@ -84,6 +84,7 @@ if __name__ == "__main__":
                                            gr.inputs.Slider(0.00000001, 0.1, default=0.001, step=0.00000001, label="Learning rate"),
                                            gr.inputs.Slider(1, 256, default=100, step=1, label="Embedding size"),
                                            gr.inputs.Slider(1, 128, default=16, step=1, label="Input Length"),
+                                           gr.inputs.Slider(1, 128, default=8, step=1, label="Convolution kernel count"),
                                            "text"],
                          outputs="text")
     iface.launch()