some initial module structuring

configs/modules.json

@@ -0,0 +1,34 @@
+{
+    "PositionalEncoder": {
+        "name": "encoder_0",
+        "inputs": [
+            "x"
+        ],
+        "meta_data": {
+            "output_dim": 8
+        }
+    },
+    "MLP": {
+        "name": "mlp_0",
+        "inputs": [
+            "encoder_0", "encoder_0"
+        ],
+        "meta_data": {
+            "num_layers": 8,
+            "num_skip_connections": [
+                3,
+                7
+            ],
+            "hidden_dim": 256,
+            "output_dim": 8
+        }
+    },
+    "RGB": {
+        "name": "rgb_0",
+        "inputs": ["mlp_0"]
+    },
+    "Density": {
+        "name": "rgb_1",
+        "inputs": ["mlp_0"]
+    }
+}

mattport/nerf/dataset/build.py

@@ -4,7 +4,7 @@ Builds the dataset
 import importlib
 import logging
 
-from mattport.nerf.dataset.base import Dataset
+from mattport.nerf.dataset.dataset import Dataset
 
 
 def build_dataset(config: dict) -> Dataset:

mattport/nerf/graph.py

@@ -3,7 +3,7 @@ The Graph module contains all trainable parameters.
 """
 from dataclasses import dataclass
 from torch import nn
-
+from mattport.nerf.modules.positional_encoder import PositionalEncoder
 
 @dataclass
 class GraphInputs:
@@ -19,10 +19,31 @@ class Graph(nn.Module):
         nn (_type_): _description_
     """
 
-    def __init__(self) -> None:
+    def __init__(self, modules_config: dict) -> None:
         super().__init__()
+
+
+        # # for key in ...
+        # temp = graph_config.positional_encoder
+        # PositionalEncoder(name=temp.name, name=temp.inputs, )
+        # self.modules = {}
+        # self.modules["name"] = MLP()
+
+
+        self.modules = {}
+
+
+        self.encoder_0 = Encoder()
+        self.mlp_0 = MLP()
+
+
         raise NotImplementedError
 
-    def forward(self):
+    def forward(self, x):
         """_summary_"""
-        raise NotImplementedError
+        # raise NotImplementedError
+        
+        for name in queue: # "encoder_0",
+            # assert name.output is not None
+            x = self.modules[name](x)
+        # self.mlp_0

mattport/nerf/modules/encoding.py

@@ -8,6 +8,10 @@ from torchtyping import TensorType
 class Encoding(nn.Module):
     """Encode an input tensor. Intended to be subclassed"""
 
+     def __init__(self, name: str, inputs: list, meta_data: dict) -> None:
+        super().__init__()
+        raise NotImplementedError
+
     def encode(self, in_tensor: TensorType[..., "input_dim"]) -> TensorType[..., "output_dim"]:
         """Encodes an input tensor.
 

mattport/nerf/modules/mlp.py

@@ -1,10 +1,4 @@
-"""
-Collection of Models
-"""
-from typing import Optional
-
 from torch import nn
-from torchtyping import TensorType
 
 
 class MLP(nn.Module):
@@ -30,7 +24,6 @@ class MLP(nn.Module):
             out_activation (Optional[nn.Module], optional): output activation function. Defaults to None.
         """
         super().__init__()
-
         layers = []
         for i in range(num_layers):
             if i == 0:

mattport/nerf/trainer.py

@@ -17,19 +17,16 @@ class Trainer(object):
         self.optimizer = None
 
     def setup_dataset(self):
-        """_summary_
-        """
+        """_summary_"""
         self.train_dataset = build_dataset(self.config.dataset)
         self.test_dataset = build_dataset(self.config.dataset)
 
     def setup_graph(self):
-        """_summary_
-        """
+        """_summary_"""
         self.graph = Graph()
 
     def setup_optimizer(self):
-        """_summary_
-        """
+        """_summary_"""
         self.optimizer = Optimizer(params=self.graph.parameters(), **self.config.optimizer)
 
     def load_checkpoint(self):