Huang Fang

Result for single model

4

Thanks for much for sharing your code, the softmap score on the Waymo leaderboard is amazing. I tried to reproduce the result, but got only 0.372 for a single model,...

WARNING:root:ONNX is running on CPU - you need to install CUDA/onnxruntime-gpu to run on GPU

1

The code I run: ``` import os import json from copy import deepcopy os.environ["BACKEND"] = "onnx" import onnxruntime as ort from unisim import TextSim ``` I got message: `WARNING:root:ONNX is...

Only one gpu is used when deploying interactive demo

3

Hi, VITA-1.5 looks great! When I run `python -m web_demo.server --model_path demo_VITA_ckpt --ip 0.0.0.0 --port 8081 `, I expect 2 gpus to be used (loading 2 models). But it appears...

Thanks for this great work! I wonder if the data used to train the monitor (classify , , ) is open-sourced? I did not find any information about the state...

OOM and slow tokenization after upgrade LLaMA-Factory

24

### Reminder - [x] I have read the above rules and searched the existing issues. ### System Info old llamafactory version: a71e6850211b6e54c07b1900cdcd7f52f7832629 new llamafactory version: cf1087d409b2513338f9f32a33b6d95919b96d91 transformers==4.51.3 ### Reproduction ```...