Qnancy

UserWarning: None of the inputs have requires_grad=True. Gradients will be None

8

由于显卡不是很充足，我将config中的freeze_vit和freeze_qformer都设置成了True，同时添加了freeze_lora，在models/videochat_mistra/videochat2_it_mistral.py中添加 ``` if self.freeze_lora: logger.info("freeze lora") for _, param in self.mistral_model.named_parameters(): param.requires_grad = False self.mistral_model = self.mistral_model.eval() self.mistral_model.train = disabled_train ``` 也就是说，我希望我只训练线性层的参数，然而在训练过程中出现 WARNING | py.warnings : .../anaconda3/envs/videochat2/lib/python3.9/site-packages/torch/utils/checkpoint.py:31: UserWarning: None of...

How to get the stage 2 checkpoint path for 3_sft.sh

5

I want to start training my own fine-tuning dataset from the stage 2 of VILA1.5-3b. I noticed in `3_sft.sh` that there is a comment for the output of the stage...

[Doc]: ValueError: Could not detect config format for no config file found.

1

### 📚 The doc issue After I `uv pip install vllm-omni`, I run `vllm serve Qwen/Qwen-Image-Edit-2509 --omni --port 8092`, but I got some eroors like: INFO 12-25 11:50:47 [__init__.py:216] Automatically...