hyperbolic-c

Before error is `can not find openssl` so solution is `sudo apt install libssl-dev`

Looking forward to your work!

Same question！Looking forward to the model based on llama3!

> 1. 是全参数 > 2. 如果是33B的话，一般需要80G显存，但通过pp并行（速度会慢），40G显存也是可以的 @guoday 你好, 我微调1.3B是两张显卡都跑到30G了, 想微调6.7B的时候显存直接爆了(batch_size=4都不行), 请问为什么消耗这么高, 好奇怪, 请教一下是要设置什么参数吗? 谢谢.

@soloice 您好,就是跑的官方例程. 请问如何修改并行策略呢谢谢大佬

> Can confirm this issue.. although it converts the model using `vocab-type hfft`, the model will not load: > > ``` > llm_load_vocab: SPM vocabulary, but newline token not found:...

Thanks for yout excellent work! Looking forward to supporting inference on ARM CPU. Further, support for [ray distributed computing ](https://github.com/vllm-project/vllm/pull/4110).

> @CISC I'm unable to test this model. I cloned the model from `[email protected]:gorilla-llm/gorilla-openfunctions-v2`. My attempt to convert with the `convert.py` script was greeted with this message: > > ```...

it looks the llamacpp support about deepseek-coder coming soon [# 5981](https://github.com/ggerganov/llama.cpp/issues/5981)

> > But when attempting to run an imatrix calculation > > Same for me with some DeepSeek-based models, which Gorilla is based on. Inference for FP16 and Q8 works,...