Handsome

run `python demo.py -p SSDLite.pth --landmarks` got error too. ``` python demo.py -p SSDLite.pth --landmarks C:\anaconda\envs\gestures2\lib\site-packages\torchvision\models\_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the...

I [modified the lib.rs code](https://github.com/FasterDecoding/REST/issues/12#issuecomment-2094677003) so that it can be applied to the qwen-7b model ([the vocab_size of the qwen-7b model is 151936](https://huggingface.co/Qwen/Qwen-7B/blob/main/config.json)), and then when I ran REST's Draft...

Hi, I tried to use your database construction method on the [qwen-7b model](https://huggingface.co/Qwen/Qwen-7B) （ The "vocab_size" of the qwen-7b model is 151936）, and then tested it with the human-eval data...

The "vocab_size" of the qwen-7b model is 151936, which exceeds the maximum value of 65536 that can be represented by 2 bytes. I have made the following modifications to the...

When building a data set for the qwen-7b model, the following error message is reported. I wonder if it will affect the construction and retrieval results of the data set?...

Hello, have you paid attention to the paper LOOKAHEAD DECODING([Break the Sequential Dependency of LLM Inference Using LOOKAHEAD DECODING](https://arxiv.org/pdf/2402.02057))? LOOKAHEAD DECODING generates draft tokens through multiple rounds of jacobi iterations...