kongds

这个的设置应该需要restart lsp-bridge才会生效，可以在`(lsp-bridge-restart-process)`之后试试

Hello, since the packages used in this repo are quite old, it seems that you are using a newer version of python with 3.11. Can you try python with 3.8.3?

Maybe gradient has exploration during training, what dataset did you use ?

The code of clustering labels is already in src/cluster.py. For example ` python ./src/cluster.py --dataset amazon670k --id $1 ` More information, please refer to #3 and #4

You need tfidf first. Please refer to https://github.com/kongds/LightXML/issues/3#issuecomment-763393334. And the cluster.py we used is directly from AttentionXML.

Hello, For OOM, It can be solve by https://github.com/kongds/LightXML/commit/0a04646535053f24608bf3ca88bc631d18f4d91c to replace the `mlb = MultiLabelBinarizer()` with `mlb = MultiLabelBinarizer(sparse_output=True)`. For the label size between 1k and 2k, i think the...

Thank you for your interest in our work. For the first question, we use CIRR and FashionIQ. You can obtain the results by running the following command (which will download...

Thank you for your interest in our work. Since I am currently checked out, I do not have access to the previously stored model weights. Additionally, I do not have...

The training details are as follows: Chat template: `\n{}\n\n` (replace {} with the training or evaluation prompt) Model: `microsoft/Phi-3-vision-128k-instruct` Other settings are the same with LLaVA, which can be found...

Since we train only the LLM of MLLM, the target modules ‘q_proj,k_proj,v_proj,o_proj,gate_proj,down_proj,up_proj’ are in the LLM. (We will remove the image encoder during training)