Spaces:
Running
on
A100
Running
on
A100
File size: 3,658 Bytes
174ae06 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 |
# Copyright (c) 2025 NVIDIA CORPORATION.
# Licensed under the MIT license.
# Adapted from https://github.com/NVlabs/VILA/tree/main under the Apache 2.0 license.
# LICENSE is in incl_licenses directory.
# Copyright 2024 NVIDIA CORPORATION & AFFILIATES
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
# SPDX-License-Identifier: Apache-2.0
import torch
import torchvision.transforms as T
from torchvision.transforms.functional import InterpolationMode
from transformers import AutoConfig, AutoModel
from transformers.image_processing_utils import BaseImageProcessor
from llava.model.multimodal_encoder.intern.configuration_intern_vit import InternVisionConfig
from llava.model.multimodal_encoder.intern.modeling_intern_vit import InternVisionModel
from llava.model.multimodal_encoder.vision_encoder import VisionTower, VisionTowerS2
def build_transform(input_size):
transform = T.Compose(
[
T.Lambda(lambda img: img.convert("RGB") if img.mode != "RGB" else img),
T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
T.ToTensor(),
T.Normalize(mean=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225)),
]
)
return transform
class InternVisionPreprocessor(BaseImageProcessor):
def __init__(self, resize_size=448):
super().__init__()
self.resize_size = resize_size
@property
def size(self):
return {"height": self.resize_size, "width": self.resize_size}
def preprocess(self, image, return_tensors):
transform = build_transform(self.resize_size)
if isinstance(image, list):
image_tensor = [transform(img) for img in image]
return {"pixel_values": image_tensor}
else:
image_tensor = transform(image)
return {"pixel_values": [image_tensor]}
class InternVisionTower(VisionTower):
def __init__(self, vision_tower, config, drop_path_rate=0.0):
super().__init__(vision_tower, config)
self._drop_path_rate = drop_path_rate
self.image_processor = InternVisionPreprocessor()
vision_config = InternVisionConfig.from_pretrained(vision_tower)
vision_config.drop_path_rate = self._drop_path_rate
self.vision_tower = InternVisionModel.from_pretrained(
vision_tower, torch_dtype=eval(config.model_dtype), config=vision_config
)
self.is_loaded = True
class InternVisionTowerS2(VisionTowerS2):
def __init__(self, vision_tower, config, drop_path_rate=0.0):
super().__init__(vision_tower, config)
self._drop_path_rate = drop_path_rate
self.image_processor = InternVisionPreprocessor(resize_size=self.scales[-1])
vision_config = InternVisionConfig.from_pretrained(vision_tower)
vision_config.drop_path_rate = self._drop_path_rate
self.vision_tower = InternVisionModel.from_pretrained(
vision_tower, torch_dtype=eval(config.model_dtype), config=vision_config
)
self.is_loaded = True
AutoConfig.register("intern_vit_6b", InternVisionConfig)
AutoModel.register(InternVisionConfig, InternVisionModel)
|