Skip to content

GitLab

Unverified Commit 367cb8ce authored by Cyrus Leung's avatar Cyrus Leung Committed by GitHub
Browse files

[Doc] [2/N] Add Fuyu E2E example for multimodal processor (#13331)

parent 54ed913f
Expand all Show whitespace changes
Inline Side-by-side
Showing with 529 additions and 31 deletions
vllm/model_executor/models/fuyu.py
View file @ 367cb8ce
...@@ -104,6 +104,8 @@ class FuyuProcessingInfo(BaseProcessingInfo): ...@@ -104,6 +104,8 @@ class FuyuProcessingInfo(BaseProcessingInfo):
image_processor = self.get_image_processor() image_processor = self.get_image_processor()
target_width = image_processor.size["width"] target_width = image_processor.size["width"]
target_height = image_processor.size["height"] target_height = image_processor.size["height"]
patch_width = image_processor.patch_size["width"]
patch_height = image_processor.patch_size["height"]
if not (image_width <= target_width and image_height <= target_height): if not (image_width <= target_width and image_height <= target_height):
height_scale_factor = target_height / image_height height_scale_factor = target_height / image_height
...@@ -113,8 +115,8 @@ class FuyuProcessingInfo(BaseProcessingInfo): ...@@ -113,8 +115,8 @@ class FuyuProcessingInfo(BaseProcessingInfo):
image_height = int(image_height * optimal_scale_factor) image_height = int(image_height * optimal_scale_factor)
image_width = int(image_width * optimal_scale_factor) image_width = int(image_width * optimal_scale_factor)
ncols = math.ceil(image_width / 30) ncols = math.ceil(image_width / patch_width)
nrows = math.ceil(image_height / 30) nrows = math.ceil(image_height / patch_height)
return ncols, nrows return ncols, nrows
def get_image_size_with_most_features(self) -> ImageSize: def get_image_size_with_most_features(self) -> ImageSize:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment