-
Hi,
First of all, thanks for your amazing work!
I try to use **potsdam dataset** on MulitMAE, my dataset architecture is like this:
> /path/to/potsdam/
> ├── train/
> │ ├── rgb/
> │ │…
-
**Suggested steps:**
* [ ] Define unsupervised learning tasks, i.e., learning tasks that don't required truth-level labels but instead relies solely on the reconstruction-level data. This is the same…
-
I was thinking about extending torchvision models with an U-Net builder for segmentation, that takes pre-trained torchvision classification models as backbone architectures in the encoder path of the …
-
1. 推理中 predict_bert_multitask 依赖训练出模型的 params.json 参数?
2. 有没有可能对,huggingface 中 llama2 不训练的模型,直接适配 predict_bert_multitask? (有一个api直接加载 pretrained_model, 生成 params.json ? )
3. 现在支持 decoder-only 的模式吗?…
-
## Description
I am currently working on a project that is built in Unity where I am modulating voices (e.g. source speech → voice modulator → target speech (elf)). I currently have an E2E flow with …
-
请教各位大佬。
文档里说:
Note that the goal of pre-training is to reconstruct the text, and the pre-trained model cannot be used for similarity calculation directly, it needs to be fine-tuned。
预训练之后的模型有一些可以…
-
Hello, I appreciate your work very much!
but I want to know how the decoder of StyleVAE pre-training model in the second stage is trained? _(vgg_normalised.pth and fc_encoder_iter_160000.pth)_
I…
-
This is really a fantastic project, thank you for your effort!
Will you provide methods to fine-tuning this model with custom dataset?
-
Just a little context -- we are using a pytorch auto-encoder based video codec, so it has 'encoder' and 'decoder' halves that work together.
We would like to only activate quantization on the decod…
-
Is there an easy way to select the sizes of the filters in the Unet-resnet?
It looks like it starts at 64, and then goes through the U with the lowest at 512. Can I change this?
It is ok if the d…