Yixin SONG

> font{ line-height: 1.6; } ul,ol{ padding-left: 20px; list-style-position: inside; } 谢谢您！我改完之后还是有四个插件无法更新，这个怎么解决呢？再次感谢您！ ***@***.*** 在2022年1月14日 ***@***.***> 写道： 编辑这个文件~/.vimplus/autoload/plug.vim，将其中所有的github.com替换为github.com.cnpmjs.org。 —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving...

same asking here :)

File "/nvme/syx/pyllama/apps/gradio/webapp.py", line 54, in load model = Transformer(model_args) File "/home/songyixin/miniconda3/envs/llama-serve/lib/python3.9/site-packages/llama/model_single.py", line 200, in __init__ self.layers.append(TransformerBlock(layer_id, params)) File "/home/songyixin/miniconda3/envs/llama-serve/lib/python3.9/site-packages/llama/model_single.py", line 168, in __init__ self.feed_forward = FeedForward( File "/home/songyixin/miniconda3/envs/llama-serve/lib/python3.9/site-packages/llama/model_single.py", line 155,...

I update the __init__.py, which switch single to parallel.py, but when I send a prompt, it will stuck ``` File "/home/songyixin/miniconda3/envs/llama-serve/lib/python3.9/site-packages/gradio/blocks.py", line 1717, in block_thread File "/home/songyixin/miniconda3/envs/llama-serve/lib/python3.9/site-packages/gradio/blocks.py", line 1524, in...

Thanks for advice. Now I can setup the service. But it will stuck. ``` File "/home/songyixin/miniconda3/envs/llama/lib/python3.9/site-packages/gradio/networking.py", line 166, in setup_tunnel File "/home/songyixin/miniconda3/envs/llama/lib/python3.9/site-packages/gradio/tunneling.py", line 95, in _start_tunnel WARNING:torch.distributed.elastic.multiprocessing.api:Sending process 196848 closing...

Thank you for your interest. The original model weight is converted from [ReLULLaMA](https://huggingface.co/SparseLLM), and the model is fine-tuned from the original llama model with just the activation function being replaced....

> It's also important to keep MoE models in mind when you expand the compatibility of PowerInfer. The ceiling for consumer grade GPUs is around 3_0 for a 8x7b so...

> Hi @YixinSong-e . I notice that you provide the ReLU-LLaMA in HF. I run the model and found that the sparsity (values lower than zero) is much lower than...

> Thanks for your reply. I have a question that in my understanding, ReGLU uses element-wise multiplication, which means those zero values after ReLU remain zero, theoretically yilelding same sparsity...

目前我们还不支持windows，我们正在解决支持windows的问题。 At present, we do not support Windows, please stay tuned. :)