YU-SHIANG HUANG

Results 3 comments of YU-SHIANG HUANG

same question here, I try to pretrain from original electra small model weights, but i get ERROR:tensorflow:Error recorded from training_loop: Restoring from checkpoint failed. This is most likely due to...

> @arthurwolf You can try building using the following, it worked for me. > > `CUDACXX=/usr/local/cuda-12/bin/nvcc CMAKE_ARGS="-DLLAMA_CUBLAS=on -DCMAKE_CUDA_ARCHITECTURES=native" FORCE_CMAKE=1 pip install llama-cpp-python --no-cache-dir --force-reinstall --upgrade` This works for me too!...

I encounter the same problem when using Mistral-7B-Instruct-v0.2. Also, I'm wondering if I need to add special tokens like [INST] [/INST] from mistral-instruct models to the implementation.