admin/lerobot
1
0
Fork 0
mirror of https://github.com/huggingface/lerobot.git synced 2026-07-19 16:02:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

remove cropping of images smaller than the crop size

Browse Source
This commit is contained in:
Bryson Jones
2025-12-15 22:20:20 -08:00
parent 25ecd16b67
commit 8a2f5aa6cb
2 changed files with 24 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/lerobot/__init__.py
+1 -1
View File
@@ -157,7 +157,7 @@ available_datasets = sorted(
) )
# lists all available policies from `lerobot/policies` # lists all available policies from `lerobot/policies`
available_policies = ["act", "multi_task_dit", "diffusion", "tdmpc", "vqbet"] available_policies = ["act", "diffusion", "tdmpc", "vqbet"]
# lists all available robots from `lerobot/robots` # lists all available robots from `lerobot/robots`
available_robots = [ available_robots = [
src/lerobot/policies/multi_task_dit/configuration_multi_task_dit.py
+23 -9
View File
@@ -14,6 +14,7 @@
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
import logging
from dataclasses import dataclass, field from dataclasses import dataclass, field
from lerobot.configs.policies import PreTrainedConfig from lerobot.configs.policies import PreTrainedConfig
@@ -144,9 +145,12 @@ class MultiTaskDiTConfig(PreTrainedConfig):
or self.image_crop_shape[1] > self.image_resize_shape[1] or self.image_crop_shape[1] > self.image_resize_shape[1]
) )
): ):
raise ValueError( logging.warning(
f"image_crop_shape {self.image_crop_shape} must be <= image_resize_shape {self.image_resize_shape}" "image_crop_shape %s must be <= image_resize_shape %s; disabling cropping.",
self.image_crop_shape,
self.image_resize_shape,
) )
self.image_crop_shape = None
# Text encoder validation # Text encoder validation
if "clip" not in self.text_encoder_name.lower(): if "clip" not in self.text_encoder_name.lower():
@@ -202,16 +206,26 @@ class MultiTaskDiTConfig(PreTrainedConfig):
def validate_features(self) -> None: def validate_features(self) -> None:
"""Validate that required input features are present and properly configured.""" """Validate that required input features are present and properly configured."""
# If the configured crop doesn't fit, disable cropping instead of erroring.
# Note: if image_resize_shape is set, cropping is applied *after* resizing.
if self.image_crop_shape is not None: if self.image_crop_shape is not None:
for key, image_ft in self.image_features.items(): for key, image_ft in self.image_features.items():
if ( # image_ft.shape is (C, H, W)
self.image_crop_shape[0] > image_ft.shape[1] effective_h, effective_w = (
or self.image_crop_shape[1] > image_ft.shape[2] self.image_resize_shape
): if self.image_resize_shape is not None
raise ValueError( else (image_ft.shape[1], image_ft.shape[2])
f"image_crop_shape {self.image_crop_shape} doesn't fit within image shape {image_ft.shape} " )
f"for '{key}'" if self.image_crop_shape[0] > effective_h or self.image_crop_shape[1] > effective_w:
logging.warning(
"image_crop_shape %s doesn't fit within effective image shape (%s, %s) for '%s'; disabling cropping.",
self.image_crop_shape,
effective_h,
effective_w,
key,
) )
self.image_crop_shape = None
break
if len(self.image_features) > 0: if len(self.image_features) > 0:
first_key, first_ft = next(iter(self.image_features.items())) first_key, first_ft = next(iter(self.image_features.items()))