jing xu comments

Results 3 comments of


                                            jing xu

Cannot find module 'tvmjs' or its corresponding type declarations EVEN has ‘npm install tvmjs’ done

> (gh_web-llm) amd00@asus00:~/llm_dev/web-llm$ npm run build（gh_web-llm） amd00@asus00：~/llm_dev/web-llm$ npm run build > > > @mlc-ai/[email protected] build@mlc-ai/[email protected] 构建 > > rollup -c 汇总 -c > > src/index.ts → lib/index.js... [!] (plugin rpt2)...

How to Reduce Model Initialization Time?

> Thanks for the question! Under the hood, weights of the model selected is downloaded from the `model_url` field (a huggingface link) in a model record:谢谢你的提问！在底层，所选模型的权重是从模型记录中的 `model_url` 字段（huggingface 链接）下载的： >...

In the Llama-2-7b-chat-hf-q4f32_1-1k model, the number of tokens in the prefill is 36 when inputting 'hello'.

> They are due to the system prompt as shown in the `mlc-chat-config.json`: https://huggingface.co/mlc-ai/Llama-2-7b-chat-hf-q4f16_1-MLC/blob/main/mlc-chat-config.json#L33-L34, which follow the specification of the official model releases.它们是由于 `mlc-chat-config.json` 中所示的系统提示造成的：https://huggingface.co/mlc-ai/Llama-2-7b-chat-hf-q4f16_1-MLC/blob/main/mlc -chat-config.json#L33-L34，遵循官方模型发布的规范。 > > If you'd...