add baseline
Browse files- src/run_medclip.py → run_medclip.py +23 -8
- src/__init__.py +0 -0
- src/__pycache__/__init__.cpython-38.pyc +0 -0
- src/__pycache__/configuration_medclip.cpython-38.pyc +0 -0
- src/__pycache__/modeling_medclip.cpython-38.pyc +0 -0
- src/__pycache__/run_medclip.cpython-38.pyc +0 -0
- src/modeling_medclip.py +1 -1
- train_model.sh +15 -0
src/run_medclip.py → run_medclip.py
RENAMED
@@ -28,6 +28,7 @@ import logging
|
|
28 |
import os
|
29 |
import sys
|
30 |
import time
|
|
|
31 |
from dataclasses import dataclass, field
|
32 |
from pathlib import Path
|
33 |
from typing import Callable, Optional
|
@@ -47,9 +48,9 @@ from flax import jax_utils
|
|
47 |
from flax.jax_utils import unreplicate
|
48 |
from flax.training import train_state
|
49 |
from flax.training.common_utils import get_metrics, shard, shard_prng_key
|
50 |
-
from
|
51 |
from transformers import AutoTokenizer, HfArgumentParser, TrainingArguments, is_tensorboard_available, set_seed
|
52 |
-
|
53 |
|
54 |
logger = logging.getLogger(__name__)
|
55 |
|
@@ -210,7 +211,6 @@ class ImageTextDataset(VisionDataset):
|
|
210 |
self,
|
211 |
root: str,
|
212 |
file_path: str,
|
213 |
-
captions_per_image=2,
|
214 |
transform: Optional[Callable] = None,
|
215 |
target_transform: Optional[Callable] = None,
|
216 |
transforms: Optional[Callable] = None,
|
@@ -224,15 +224,21 @@ class ImageTextDataset(VisionDataset):
|
|
224 |
self.image_paths = []
|
225 |
|
226 |
for example in examples:
|
227 |
-
self.captions.
|
228 |
-
self.image_paths.
|
229 |
|
230 |
def _load_image(self, idx: int):
|
231 |
path = self.image_paths[idx]
|
232 |
return read_image(path, mode=ImageReadMode.RGB)
|
233 |
|
234 |
def _load_target(self, idx):
|
235 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
236 |
|
237 |
def __getitem__(self, index: int):
|
238 |
image = self._load_image(index)
|
@@ -290,6 +296,17 @@ def main():
|
|
290 |
else:
|
291 |
model_args, data_args, training_args = parser.parse_args_into_dataclasses()
|
292 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
293 |
if (
|
294 |
os.path.exists(training_args.output_dir)
|
295 |
and os.listdir(training_args.output_dir)
|
@@ -351,14 +368,12 @@ def main():
|
|
351 |
train_dataset = ImageTextDataset(
|
352 |
data_args.data_dir,
|
353 |
data_args.train_file,
|
354 |
-
captions_per_image=2,
|
355 |
transform=preprocess,
|
356 |
)
|
357 |
|
358 |
eval_dataset = ImageTextDataset(
|
359 |
data_args.data_dir,
|
360 |
data_args.validation_file,
|
361 |
-
captions_per_image=1,
|
362 |
transform=preprocess,
|
363 |
)
|
364 |
|
|
|
28 |
import os
|
29 |
import sys
|
30 |
import time
|
31 |
+
import getpass
|
32 |
from dataclasses import dataclass, field
|
33 |
from pathlib import Path
|
34 |
from typing import Callable, Optional
|
|
|
48 |
from flax.jax_utils import unreplicate
|
49 |
from flax.training import train_state
|
50 |
from flax.training.common_utils import get_metrics, shard, shard_prng_key
|
51 |
+
from src.modeling_medclip import FlaxHybridCLIP
|
52 |
from transformers import AutoTokenizer, HfArgumentParser, TrainingArguments, is_tensorboard_available, set_seed
|
53 |
+
import wandb
|
54 |
|
55 |
logger = logging.getLogger(__name__)
|
56 |
|
|
|
211 |
self,
|
212 |
root: str,
|
213 |
file_path: str,
|
|
|
214 |
transform: Optional[Callable] = None,
|
215 |
target_transform: Optional[Callable] = None,
|
216 |
transforms: Optional[Callable] = None,
|
|
|
224 |
self.image_paths = []
|
225 |
|
226 |
for example in examples:
|
227 |
+
self.captions.append(example["caption"])
|
228 |
+
self.image_paths.append(f'{root}/{example["image_path"]}')
|
229 |
|
230 |
def _load_image(self, idx: int):
|
231 |
path = self.image_paths[idx]
|
232 |
return read_image(path, mode=ImageReadMode.RGB)
|
233 |
|
234 |
def _load_target(self, idx):
|
235 |
+
sections = self.captions[idx]
|
236 |
+
longest_section = max(
|
237 |
+
filter(lambda x: isinstance(x, str), sections.values()),
|
238 |
+
key=len
|
239 |
+
)
|
240 |
+
|
241 |
+
return longest_section
|
242 |
|
243 |
def __getitem__(self, index: int):
|
244 |
image = self._load_image(index)
|
|
|
296 |
else:
|
297 |
model_args, data_args, training_args = parser.parse_args_into_dataclasses()
|
298 |
|
299 |
+
if jax.process_index() == 0:
|
300 |
+
wandb.init(
|
301 |
+
entity=getpass.getuser(),
|
302 |
+
project='medclip',
|
303 |
+
sync_tensorboard=True
|
304 |
+
)
|
305 |
+
|
306 |
+
wandb.config.update(model_args)
|
307 |
+
wandb.config.update(data_args)
|
308 |
+
wandb.config.update(training_args)
|
309 |
+
|
310 |
if (
|
311 |
os.path.exists(training_args.output_dir)
|
312 |
and os.listdir(training_args.output_dir)
|
|
|
368 |
train_dataset = ImageTextDataset(
|
369 |
data_args.data_dir,
|
370 |
data_args.train_file,
|
|
|
371 |
transform=preprocess,
|
372 |
)
|
373 |
|
374 |
eval_dataset = ImageTextDataset(
|
375 |
data_args.data_dir,
|
376 |
data_args.validation_file,
|
|
|
377 |
transform=preprocess,
|
378 |
)
|
379 |
|
src/__init__.py
ADDED
File without changes
|
src/__pycache__/__init__.cpython-38.pyc
ADDED
Binary file (128 Bytes). View file
|
|
src/__pycache__/configuration_medclip.cpython-38.pyc
ADDED
Binary file (4.17 kB). View file
|
|
src/__pycache__/modeling_medclip.cpython-38.pyc
ADDED
Binary file (12.9 kB). View file
|
|
src/__pycache__/run_medclip.cpython-38.pyc
ADDED
Binary file (16.8 kB). View file
|
|
src/modeling_medclip.py
CHANGED
@@ -18,7 +18,7 @@ from typing import Optional, Tuple
|
|
18 |
import flax.linen as nn
|
19 |
import jax
|
20 |
import jax.numpy as jnp
|
21 |
-
from
|
22 |
from flax.core.frozen_dict import FrozenDict
|
23 |
from transformers import FLAX_MODEL_MAPPING, FlaxCLIPVisionModel
|
24 |
from transformers.modeling_flax_utils import FlaxPreTrainedModel
|
|
|
18 |
import flax.linen as nn
|
19 |
import jax
|
20 |
import jax.numpy as jnp
|
21 |
+
from src.configuration_medclip import HybridCLIPConfig
|
22 |
from flax.core.frozen_dict import FrozenDict
|
23 |
from transformers import FLAX_MODEL_MAPPING, FlaxCLIPVisionModel
|
24 |
from transformers.modeling_flax_utils import FlaxPreTrainedModel
|
train_model.sh
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
python run_medclip.py \
|
2 |
+
--output_dir model \
|
3 |
+
--text_model_name_or_path="allenai/scibert_scivocab_uncased" \
|
4 |
+
--vision_model_name_or_path="openai/clip-vit-base-patch32" \
|
5 |
+
--tokenizer_name="allenai/scibert_scivocab_uncased" \
|
6 |
+
--data_dir="/home/shared/data/mimic-cxr" \
|
7 |
+
--train_file="/home/shared/data/mimic-cxr/train_dataset.json" \
|
8 |
+
--validation_file="/home/shared/data/mimic-cxr/validate_dataset.json" \
|
9 |
+
--do_train --do_eval \
|
10 |
+
--num_train_epochs="40" --max_seq_length 512 \
|
11 |
+
--per_device_train_batch_size="64" \
|
12 |
+
--per_device_eval_batch_size="64" \
|
13 |
+
--learning_rate="5e-5" --warmup_steps="0" --weight_decay 0.1 \
|
14 |
+
--overwrite_output_dir \
|
15 |
+
--preprocessing_num_workers 32 \
|