|
324c690841
|
model and datasets adapted to new yolo format. reinforced learning is now implemented. correct_bboxes.py now capable of exporting data to train file
|
2023-11-16 12:19:57 +01:00 |
|
|
9dcf54c139
|
adapted data structure
|
2023-11-07 15:58:10 +01:00 |
|
|
323d79b27a
|
bf
|
2023-10-27 11:27:31 +02:00 |
|
|
f3e88078c4
|
bf in InferenceDataset in datasets.py
|
2023-10-27 11:25:32 +02:00 |
|
|
873758d461
|
embed for bf
|
2023-10-27 11:22:59 +02:00 |
|
|
cbbf839e72
|
bf
|
2023-10-27 11:16:51 +02:00 |
|
|
3690907f93
|
inference rework
|
2023-10-27 11:15:53 +02:00 |
|
|
a91e61f514
|
stupid bug fixed
|
2023-10-26 12:51:45 +02:00 |
|
|
5898a885fd
|
bf
|
2023-10-26 12:50:50 +02:00 |
|
|
030dcfab43
|
changed TRAIN_DIR to DATA_DIR in all files since it is more representative in this code. test and train datasets are defined by csv files
|
2023-10-26 12:47:29 +02:00 |
|
|
0adfc22ec4
|
check if i can leave image name in target as return from dataset ... to(DEVICE) problematic
|
2023-10-26 09:25:02 +02:00 |
|
|
4373b7c453
|
bug fixed ... no double images any longer... repeat trainging because there we have the same issue...
|
2023-10-26 09:00:38 +02:00 |
|
|
9905b6039e
|
big bug... not unique imaged :D
|
2023-10-26 08:56:43 +02:00 |
|
|
b89cf321f9
|
bf
|
2023-10-26 08:40:24 +02:00 |
|
|
b1e5b9e665
|
inference updated
|
2023-10-26 08:38:14 +02:00 |
|
|
e1a97ac493
|
train test split implemented using different csv files
|
2023-10-25 15:31:00 +02:00 |
|
|
8a6e7df57f
|
generated huge dataset
|
2023-10-24 15:15:19 +02:00 |
|
|
30a9e71e76
|
training loop runs ... no feed it
|
2023-10-23 15:26:39 +02:00 |
|
|
3f66b4a39a
|
alot of work. dataset and loader not created
|
2023-10-23 14:43:01 +02:00 |
|