Yung-Sung Chuang

@sheetalsh456 You're welcome!

I've done all the setups suggested by @wookayin but still getting this error: `ld: library not found for -lmujoco210` I've also tried the following possible steps: ``` ❯ export LIBRARY_PATH=$LIBRARY_PATH:/Users/$USER/.mujoco...

I think both implementations should be fine. In theory, we should not apply `.mean()` because `.sum()` is the true estimation of the probability of the whole sentence: $P(x_1, x_2, ...,...

Hi @naknak-choi Thanks for the issue! I had fixed this bug and pushed it in the latest commit. Sorry that I forgot to remove the unused model args. The current...

Hi @VityaVitalich Thanks for your interests in our project! The `teacher_forcing_seq` was used when we had a given sequence for calculating the lookback ratio. The given sequence may not be...

Hi @nikilpatel94 Can you use python 3.9.5 and cuda toolkit 11.7? I believe you will be able to install torch 2.0.0 in this setting. Thanks!

Hi! It's from this part of the code: https://github.com/voidism/Lookback-Lens/blob/8298a9525943b3398fe20316867a53b322647c98/transformers-4.32.0/src/transformers/generation/stopping_criteria.py#L43 If you `pip install ./transformers-4.32.0`, you should be able to import it, at least it worked in my test. Let my...

Hi @OliverLeeXZ DoLa was supported by Huggingface transformers last year: https://huggingface.co/docs/transformers/main/generation_strategies#dola-decoding However, we didn't test the decoding results on the latest models like Qwen, Llama3, etc. For reproducing the official...