Xiao

get操作后，是不是要将key的指针移到最前面？？

这不是北航和MSRA的合作吗？发了好几年了，我感觉应该不会开源发自我的iPhone------------------ Original ------------------From: Jack Chen ***@***.***>Date: Fri,Mar 18,2022 10:11 AMTo: dyweb/papers-notebook ***@***.***>Cc: Subscribed ***@***.***>Subject: Re: [dyweb/papers-notebook] Gandiva: Introspective Cluster Schedulingfor Deep Learning (#109) 这篇论文是北邮的博士学长的一作，据说有开源的计划，可以期待一下。 论文是基于这样几个 key insight: 深度学习是反馈驱动的探索，用户经常运行一批训练，取其中结果最好的。这个可以理解为是类似参数搜索，模型结构搜索这样的场景。 在资源使用的异构性，导致很难得到最优解 intra-job...

another problem. ``` new_nodes = get_retrieved_nodes( "What is Lyft's response to COVID-19?", vector_top_k=10, reranker_top_n=5, with_reranker=True, ) visualize_retrieved_nodes(new_nodes) ``` the error is ``` IndexError Traceback (most recent call last) Cell In[12],...

@dosubot

thanks, I fix the error.

my code is below ``` import nest_asyncio nest_asyncio.apply() import logging import sys logging.basicConfig(stream=sys.stdout, level=logging.INFO) logging.getLogger().addHandler(logging.StreamHandler(stream=sys.stdout)) from llama_index.core import VectorStoreIndex, SimpleDirectoryReader from llama_index.core.postprocessor import LLMRerank from IPython.display import Markdown, display import...

Also, if I want to run query on the rerank . how?

> Currently, a config file is not necessary. See this as the latest CAI GPT example https://github.com/hpcaitech/ColossalAI/blob/main/examples/language/gpt/README.md I can't find the [import model_zoo.gpt.gpt as col_gpt](https://github.com/hpcaitech/ColossalAI-Examples/blob/main/language/gpt/model/pipeline_gpt1d.py#L12).

> Override the LLM class complete() and/ chat() function, insert a timer and call super? good idea.

> Yes, your approach to measure the end-to-end time for each single new query generated in a multi-step query is a good one. You can indeed override the `complete()` and...