InternImage

InternImage copied to clipboard

[bug] how to train fine-tuning classification model (size mismatch for head.bias: copying a param with shape torch.Size([1000]) from checkpoint, the shape in current model is torch.Size([4]).)

4

I customized the config.py. how to train fine-tuning classification model?

jeonga0303

Hello, As I see in the file `InternImage/segmentation/mmseg_custom/models/decode_heads /mask2former_head.py`, there is this line: `self.cls_embed = nn.Linear(feat_channels, self.num_classes + 1)` I am wondering why the `+1` is there. As far as...

AndreiMihalea

In InternImage there is no publicly available code for single image classification inference. Could someone please provide an inference script or command for this purpose? I would be very grateful.

unclelee97

train on custom dataset

1

How can I train instance segmentation on my own dataset? Should I use this link "https://github.com/OpenGVLab/InternImage/tree/master/detection"? I didn't find a folder for instance segmentation like the folders for segmentation and...

Homa2018

Training on custom datasets

6

Hi, great work! I want to train detection on other datasets, but it seems that your config files only support training on COCO. How can I train your model on...

quanhaol

Hi, thanks for the amazing work. My application uses 3D image dataset, I want to use DCNv4 layer block. However, the current implementation supports 2D input, how can we customize...

mfaizan-ai

Hello, I was looking into the released models, and it looks like the checkpoint for object detection based on OpenImagesV6 is not released? I was wondering if you plan to...

TouqeerAhmad

I built an environment containing all necessary library but when I run this code to sgement images, this error occurs. here is my pip list: Package Version ------------------------ ------------ addict...

kll-df

delete meaningless dot

ArtificialZeng

KeyError: "EncoderDecoder: 'InternImage is not in the models registry'"

1

Hey! I have been running the pre-trained Internimage-XL fine for a while, but now I'm unable to run it. The code I'm using has not changed and did not give...

jusba