NielsRogge

@xieenze this is a notebook illustrating fine-tuning SegFormer on custom data: https://colab.research.google.com/drive/15JeOp3KxEjeTxG74DZc1cGZVaT5gPQjj?usp=sharing However, I'm not sure whether it works properly already (loss is going down nicely, but inference results don't...

Thanks for looking into it. Is this part wrong? ``` if self.reduce_zero_label: if segmentation_maps is not None: for idx, map in enumerate(segmentation_maps): if not isinstance(map, np.ndarray): map = np.array(map) #...

Yes that's the case, as can be seen [here](https://github.com/NielsRogge/transformers/blob/ffdc58af4fb9b611c48e2acad8a999b8a9863713/src/transformers/models/segformer/modeling_segformer.py#L780). So then I'm mostly ready. However, it's weird that the inference part doesn't look as expected, after fine-tuning the model. Or...

> The "seg_pad_val" is used in class Pad, but the "ignore_index" is not used in loss "CrossEntropyLoss" during training, so does it really ignore the ignore_index during training? I'm not...

I'm have an own implementation, I'm not using the mmseg code base.

Hi, Yeah I've tested it on a small dataset and it seems to work well. I'm currently working on another model, but once I have the time I'll add SegFormer...

Hi folks (cc @jpWang), Happy to share that I've got a working HuggingFace implementation. I'm able to reproduce the results (F1 of 88% on FUNSD). The model is here: https://huggingface.co/nielsr/lilt-roberta-en-base-finetuned-funsd...

Hi @jpWang, I've opened a PR here to add LiLT to 🤗 Transformers: https://github.com/huggingface/transformers/pull/19450. Would you like to create an organization for the company/research institute you're part of, such that...

Hi, What's the URL of the HuggingFace organization? I can't seem to find SCUT-DLVCLab on [hf.co](https://huggingface.co/).

Thanks a lot! Now 2 checkpoints are transferred there. Could you give me your email address, such that I can set up a Slack channel? I'd like to discuss some...