264 Commits

Author SHA1 Message Date
6318d52f12 Use const T* for input tensors in layer and tensor operations 2025-11-28 21:41:38 +01:00
71dc5a924d Move dtype size implementation to cpp file 2025-11-27 23:29:54 +01:00
7e27c87673 Fix compilation errors and warnings 2025-11-27 22:41:49 +01:00
e79667671a Refactor size calculations in layers and backend 2025-11-27 22:01:09 +01:00
c855ae89ec Refactor Tensor methods to use void* for data handling and add device_ptr method 2025-11-27 21:18:51 +01:00
9ff214d759 Refactor CUDA kernels and tensor operations for type generality 2025-11-26 20:47:55 +01:00
13d3d38b68 Add dtype parameter to layer constructors 2025-11-26 00:19:33 +01:00
84153ac49c Add default dtype to backend 2025-11-25 23:42:19 +01:00
ad079560ff Update CMakeLists.txt 2025-11-25 19:08:55 +01:00
60964cf294 Move factory implementation out of header 2025-11-24 22:01:54 +01:00
a40ba96d4f Implement backend factory 2025-11-24 21:53:47 +01:00
a97ff8e1f6 Update main include file 2025-11-23 21:07:34 +01:00
38cb0c9ac0 Restructure include paths 2025-11-23 20:57:08 +01:00
4161caf3e1 Update BatchNorm2d to return sizes for running mean and var 2025-11-23 20:48:41 +01:00
9f1a56c699 Refactor Layer interface to return size of weights and biases instead of Tensor references 2025-11-23 20:44:25 +01:00
547cd0c224 Remove unnecessary inclusion of cuda_helper.cuh in pool.cu 2025-11-23 19:21:22 +01:00
1102aef293 Implement custom Shape struct with __device__ support 2025-11-23 19:21:06 +01:00
82a0e7c19d Fix some compilation errors 2025-11-23 18:50:57 +01:00
51bcee01ab Migrate model class to Tensor 2025-11-22 22:40:38 +01:00
ca44ea4436 Migrate module to tensors 2025-11-22 18:02:42 +01:00
104d6ea33d Fix small layer issues 2025-11-22 00:33:51 +01:00
4c8b2ef537 Migrate add layer to tensors 2025-11-22 00:12:20 +01:00
aeb1739c46 Migrate concat layer 2025-11-21 23:52:58 +01:00
fd4775faa4 Migrate batch norm layer 2025-11-21 23:24:14 +01:00
5679dc0a50 Add avgPool2d implementation 2025-11-21 19:39:30 +01:00
c83e1f0c45 Implement InvalidShapeException 2025-11-21 18:54:45 +01:00
6685aa6629 WIP Migrate AvgPool2d 2025-11-19 23:21:18 +01:00
e4d05931d4 Migrate MaxPool2d layer to Tensors 2025-11-19 21:44:19 +01:00
7896ff0e24 Remove const from layer input 2025-11-19 20:37:41 +01:00
dfdfa19022 Migrate conv2d layer to Tensor 2025-11-19 20:20:46 +01:00
10c84d75fc Fix Tensor issues 2025-11-18 22:38:56 +01:00
4c26efe826 Fix some dense layer issues 2025-11-18 22:17:08 +01:00
7f203b8947 WIP Migrate Dense layer 2025-11-18 21:12:47 +01:00
64eac7050b WIP Migrate Dense layer 2025-11-18 19:33:51 +01:00
24606491a3 WIP Refactor Layer and Activation classes 2025-11-18 19:10:18 +01:00
6340b27055 Refactor Backend and Layer interfaces 2025-11-18 18:27:57 +01:00
25670f90c4 Migrate Activation layer to Tensor 2025-11-17 22:51:54 +01:00
d231e515b1 WIP Implement Tensor constructor and destructor 2025-11-17 22:36:29 +01:00
6744c8964f WIP Migrate vector utils to Tesnor 2025-11-17 22:15:27 +01:00
6133fb20af WIP Migrate Activation to Tensor 2025-11-16 22:17:46 +01:00
64bf9197ff Add simple Tensor class 2025-11-16 19:31:09 +01:00
98ad4ac760 Migrate conv2d layer 2024-09-10 21:48:47 +02:00
757584544c Migrate conv2d layer 2024-09-10 21:19:50 +02:00
74f49d6a00 Migrate output layer 2024-09-10 19:20:00 +02:00
f7b525e494 Migrate input layer 2024-09-10 19:11:21 +02:00
fe7c16ac36 Migrate concat layer 2024-09-09 22:16:22 +02:00
a0665fb05c Migrate max pooling 2024-09-09 21:48:36 +02:00
8a5d556b7e Rename cpp version of avg pooling 2024-09-09 21:37:00 +02:00
76e5225001 Migrate avg pooling 2024-09-09 21:36:13 +02:00
75475790ac Migrate Dense layer 2024-09-08 13:36:53 +02:00