Naseem Machlovi

I am facing the same issue ..I have tried all the solutions. Can anyone help me

> I have created a [tiny HF space](https://zeel-twro.hf.space/) for everyone to see how `train_on_response_only` will work on their models and inputs. You can also copy and share the URL populated...

> > [@patel-zeel](https://github.com/patel-zeel) Thank you for sharing!I got the idea that it masks the inputs in chat format. What I am confused about is when I am finetuning a Phi4...

Yes, the current files don't support any kind of video training. I have been digging into this for the last 48 hours, and every time, there are errors, simply because...

Hi, there. I have been able to determine the training process for the model. You can refer to my issue here, issue #318. Hope this works for you. The overall...

> What's your video format? The original training configuration is full of COCO data loaders. I assume we need to change those as well? Below is my format, try to...

> ### Practical steps to fine-tune > 1. Choose whether you train on images or videos. > 2. Prepare media: > > * Images: place images in a folder and...

**1. Place it under the utils files. Depending upon the model (image/video), you may have to update the wrapper.** sam3/train/utils/freeze_utils.py ``` import torch def unwrap_model(model): # Unwrap DDP/DataParallel if hasattr(model,...

> > 我想在自己的数据集上微调模型。但我需要对特定层次进行训练。 > > 我想知道有没有可以在 yaml 配置文件中更改或添加属性，来冻结不同层的参数？ > > Hello, may I ask whether you have resolved the issues related to fine-tuning SAM3?I am currently trying to fine-tune SAM3...

Please follow this issue for JSON formatting. https://github.com/facebookresearch/sam3/issues/163#issuecomment-3558320492 Image: ``` { "info": {"description": "example image dataset"}, "images": [ {"id": 1, "file_name": "images/img1.jpg", "width": 1280, "height": 720} ], "annotations": [ {...