| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2024-07-16 | add: backward cblas optimization done | jvech | |
| weights update for loop was changed for a cblas routine | |||
| 2023-10-21 | setup: man page improved | jvech | |
| Documentation: Example sections was improved Files sections was added Fixes: default config filepath was fixed | |||
| 2023-10-21 | setup: Install and uninstall rules added | jvech | |
| The man page must have a section specifying where to save configurations files for loading the network architecture and weights. | |||
| 2023-09-11 | add: static file reading implemented | jvech | |
| Now is possible to read data from standard input | |||
| 2023-09-03 | doc: LICENSE and some documentation added | jvech | |
| 2023-09-03 | fix: write network parameter fixed | jvech | |
| 2023-08-30 | add: network read and write done json_read reactored | jvech | |
| 2023-08-27 | add: data and network initialization done | jvech | |
| train subcommand can read and train the network TODO: - Refactor json_read() to parse multiple labels - Implement a function to save network weights once the network have trained - Implement a function to load trained weights to use with predict subcommand | |||
| 2023-08-26 | fix: cfg syntax verification and usage messages improved | jvech | |
| 2023-08-25 | add: layer architecture loading from cfg done | jvech | |
| 2023-08-23 | add: basic configuration parsing implemented | jvech | |
| parameters implemented: - input_keys - label_keys - epochs - labels - weights_path Network weights and architecture loading is not implemented yet. | |||
| 2023-08-17 | [ADD]: Cli implemented in a different function | jvech | |
| The program config parameters where encapsulated in a struct to keep track of them without implementing a lot of functions | |||
| 2023-08-16 | add: Cli defined | jvech | |
| 2023-08-14 | add: cli interface defined | jvech | |
| 2023-08-12 | add: softplus activation added | jvech | |
| 2023-08-12 | add: nn_network_predict done | jvech | |
| 2023-08-08 | add: pass activations to a separate file | jvech | |
| The point is have this file for new activation function without bloating nn.c file | |||
| 2023-08-08 | add: activation and cost functions added | jvech | |
| Training function were fixed and new activation and cost function were implemented | |||
| 2023-08-06 | add: nn_network_train() added | jvech | |
| The function needs to be tested | |||
| 2023-08-06 | refactor: Activation and Layer struct refactored | jvech | |
| 2023-08-05 | add: nn_backward() calloc error handling added | jvech | |
| 2023-08-05 | add: bias backward propagation implemented | jvech | |
| 2023-08-04 | add: nn_backward implemented | jvech | |
| It needs to be tested and some backward layer functions were redefined to improve readability | |||
| 2023-08-02 | add: layer backward done | jvech | |
| 2023-08-02 | add: delta calculation for hidden and output layer done | jvech | |
| 2023-07-29 | add: nn_layer_forward splited | jvech | |
| The function nn_layer_forward was changed from f(X*W) to X*W and nn_layer_map_activation will map the activation function through the output of nn_layer_forward. | |||
| 2023-07-28 | add: network forward pass implemented | jvech | |
| 2023-07-27 | fix: weights allocation mistake fixed | jvech | |
| 2023-07-27 | fix: random weights generator bug fixed | jvech | |
| 2023-07-24 | doc: backpropagation notes added | jvech | |
| 2023-07-13 | add: json_read improved | jvech | |
| 2023-07-13 | add: forward pass with openblas in progress | jvech | |
| 2023-07-12 | add: weights initialization and deallocation done | jvech | |
| 2023-07-10 | add: json file function implemented | jvech | |
 |
index : ml | |
| A multilayer perceptron network written in C |
| aboutsummaryrefslogtreecommitdiff |