Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Bugfix] Fix precision error in LLaVA-NeXT feature size calculation #11735

Merged
merged 1 commit into from
Jan 4, 2025
Merged

[Bugfix] Fix precision error in LLaVA-NeXT feature size calculation #11735

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 1 addition & 2 deletions tests/models/decoder_only/vision_language/processing/test_llava_next.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,10 +15,9 @@ def processor_for_llava_next():
return LlavaNextMultiModalProcessor


# FIXME: image_size [(198, 176), (176, 198)]
@pytest.mark.parametrize("model_id", ["llava-hf/llava-v1.6-mistral-7b-hf"])
@pytest.mark.parametrize("image_size", [(1669, 2560), (2560, 1669), (183, 488),
(488, 183)])
(488, 183), (198, 176), (176, 198)])
@pytest.mark.parametrize("num_imgs", [1, 2])
def test_processor_prompt_replacements(
processor_for_llava_next,
Expand Down
14 changes: 10 additions & 4 deletions vllm/model_executor/models/llava_next.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@
from typing import (Iterable, List, Literal, Mapping, Optional, Set, Tuple,
TypedDict, Union)

import numpy as np
import torch
import torch.nn as nn
from transformers import BatchFeature, LlavaNextConfig, LlavaNextProcessor
Expand Down Expand Up @@ -139,16 +140,21 @@ def _get_num_unpadded_features(
current_height = npatches * num_patch_height
current_width = npatches * num_patch_width

original_aspect_ratio = original_width / original_height
current_aspect_ratio = current_width / current_height
# NOTE: HF resizes based on float32
original_aspect_ratio = np.array(original_width / original_height,
dtype=np.float32)
current_aspect_ratio = np.array(current_width / current_height,
dtype=np.float32)

if original_aspect_ratio > current_aspect_ratio:
scale_factor = current_width / original_width
scale_factor = np.array(current_width / original_width,
dtype=np.float32)
new_height = int(original_height * scale_factor)
padding = (current_height - new_height) // 2
current_height -= 2 * padding
else:
scale_factor = current_height / original_height
scale_factor = np.array(current_height / original_height,
dtype=np.float32)
new_width = int(original_width * scale_factor)
padding = (current_width - new_width) // 2
current_width -= 2 * padding
Expand Down
23 changes: 15 additions & 8 deletions vllm/model_executor/models/llava_onevision.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@
from typing import (Iterable, List, Literal, Mapping, Optional, Set, Tuple,
TypedDict, Union)

import numpy as np
import torch
import torch.nn as nn
from transformers import (BatchFeature, LlavaOnevisionConfig,
Expand Down Expand Up @@ -127,18 +128,24 @@ def _get_num_unpadded_features(
current_height = npatches * num_patch_height
current_width = npatches * num_patch_width

original_aspect_ratio = original_width / original_height
current_aspect_ratio = current_width / current_height
# NOTE: HF resizes based on float32
original_aspect_ratio = np.array(original_width / original_height,
dtype=np.float32)
current_aspect_ratio = np.array(current_width / current_height,
dtype=np.float32)

if original_aspect_ratio > current_aspect_ratio:
new_height = int(original_height *
(current_width / original_width))
scale_factor = np.array(current_width / original_width,
dtype=np.float32)
new_height = int(original_height * scale_factor)
padding = (current_height - new_height) // 2
current_height -= padding * 2
current_height -= 2 * padding
else:
new_width = int(original_width *
(current_height / original_height))
scale_factor = np.array(current_height / original_height,
dtype=np.float32)
new_width = int(original_width * scale_factor)
padding = (current_width - new_width) // 2
current_width -= padding * 2
current_width -= 2 * padding

unpadded_features = current_height * current_width
newline_features = current_height
Expand Down
Loading