Change where images weres stored
Browse files- .gitignore +1 -0
- src/config.py +2 -2
- src/data.py +3 -2
.gitignore
CHANGED
@@ -1,3 +1,4 @@
|
|
1 |
.DS_Store
|
2 |
.vscode/
|
3 |
pyrightconfig.json
|
|
|
|
1 |
.DS_Store
|
2 |
.vscode/
|
3 |
pyrightconfig.json
|
4 |
+
*.jpg
|
src/config.py
CHANGED
@@ -4,7 +4,7 @@ import pydantic
|
|
4 |
|
5 |
MAX_DOWNLOAD_TIME = 0.2
|
6 |
|
7 |
-
IMAGE_DOWNLOAD_PATH = pathlib.Path("/
|
8 |
|
9 |
|
10 |
class DataConfig(pydantic.BaseModel):
|
@@ -30,7 +30,7 @@ class ModelConfig(pydantic.BaseModel):
|
|
30 |
|
31 |
class TrainerConfig(pydantic.BaseModel):
|
32 |
epochs: int = 20
|
33 |
-
batch_size: int =
|
34 |
learning_rate: float = 5e-4
|
35 |
accumulate_grad_batches: int = 1
|
36 |
temperature: float = 1.0
|
|
|
4 |
|
5 |
MAX_DOWNLOAD_TIME = 0.2
|
6 |
|
7 |
+
IMAGE_DOWNLOAD_PATH = pathlib.Path("./data/images")
|
8 |
|
9 |
|
10 |
class DataConfig(pydantic.BaseModel):
|
|
|
30 |
|
31 |
class TrainerConfig(pydantic.BaseModel):
|
32 |
epochs: int = 20
|
33 |
+
batch_size: int = 64
|
34 |
learning_rate: float = 5e-4
|
35 |
accumulate_grad_batches: int = 1
|
36 |
temperature: float = 1.0
|
src/data.py
CHANGED
@@ -102,7 +102,8 @@ if __name__ == "__main__":
|
|
102 |
)
|
103 |
train_dl, valid_dl = get_dataset(transform, tokenizer, hyper_parameters)
|
104 |
|
|
|
|
|
|
|
105 |
for batch in tqdm(train_dl):
|
106 |
continue
|
107 |
-
|
108 |
-
print("hellow")
|
|
|
102 |
)
|
103 |
train_dl, valid_dl = get_dataset(transform, tokenizer, hyper_parameters)
|
104 |
|
105 |
+
batch = next(iter(train_dl))
|
106 |
+
print({k: v.shape for k, v in batch.items()}) # torch.Size([1, 3, 128, 128])
|
107 |
+
|
108 |
for batch in tqdm(train_dl):
|
109 |
continue
|
|
|
|