add support of baichuan-7b

[foldl]

This PR adds support of baichuan-7b model.

baichuan-7B is an open-source large-scale pre-trained model developed by Baichuan Intelligent Technology. Based on the Transformer architecture, it is a model with 7 billion parameters trained on approximately 1.2 trillion tokens. It supports both Chinese and English, with a context window length of 4096. It achieves the best performance of its size on standard Chinese and English authoritative benchmarks (C-EVAL/MMLU).

This model doubles the vocabulary (64k) of LLaMa, which seems interesting.

[liaoweiguo]

now there is a 13b version, seems more powerful

[dillfrescott]

I quantized the 13B to 8_0 and it doesnt seem to work with this yet

[foldl]

13B model is changed (ALiBi position embedding). So it can't be supported out of the box like 7B.

[LiuKai22]

13B model is changed (ALiBi position embedding). So it can't be supported out of the box like 7B.

ggml already supports ALiBi OP, you just need to add some changes in llama.cpp's "llama_eval_internal" and replace RoPE with ALiBi.

But ggml only supports CPU inference for Baichuan 13B since CUDA version of alibi OP is not released, like "ggml_cuda_op_alibi" and "alibi_f32_cuda" in ggml-cuda.cu.
So, there is no GPU inference acceleration for models like Baichuan 13B. Is anyone has a plan on this? @ggerganov @foldl @JohannesGaessler

[dereklll]

13B model is changed (ALiBi position embedding). So it can't be supported out of the box like 7B.

ggml already supports ALiBi OP, you just need to add some changes in llama.cpp's "llama_eval_internal" and replace RoPE with ALiBi.

But ggml only supports CPU inference for Baichuan 13B since CUDA version of alibi OP is not released, like "ggml_cuda_op_alibi" and "alibi_f32_cuda" in ggml-cuda.cu. So, there is no GPU inference acceleration for models like Baichuan 13B. Is anyone has a plan on this? @ggerganov @foldl @JohannesGaessler

same.