Aru Sharma

Results 81 comments of Aru Sharma

Hi @juntao I was playing around different methods of creating a knowledge base as mentioned [here](https://docs.gaianet.ai/category/knowledge-bases) and I started with the pdf one using llamaparse and used llama 3.1 7B...

Hi @juntao I tried creating a summary of reports here is a [notebook](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Code/llama3_1_summarizer.ipynb) that does the same. The results aren't very good but they can be improved by using different...

Hi @juntao Update of summarization I tried summarization by llama 3 using gaianet. I wrote python [script](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Code/pdf_preprocess.py) to chunk the 10-Q pdf page wise and then create a csv [file](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Dataset/CSV/test.csv)...

Hi @juntao [This](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Dataset/PDF/Document.pdf) is the apple's 10-Q filing of 2017 and [this](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Dataset/CSV/out_summary.csv) is the final csv containing the page wise summary along with the context. [This](https://huggingface.co/datasets/FiendHunter/Financial_chatbot/resolve/main/summary_qna.tar.gz) is the snapshot from...

Greetings @juntao Using [this](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Code/3.1%20qna_generate.py) and [this](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Code/3.2%20qna_format.py) I was able to generate [this](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Dataset/Text/test.txt) which can be fed directly to finetune a llama and improve the responses for our bot. Can you...

Hi @juntao I created this [text](https://github.com/staru09/LFX_test/blob/main/Financial_chatbot/Dataset/Text/qna.txt) file containing qna related to 10-Q filings of 2017 using gaianet but I feel in order to finetune models for being a better financial...

Hi @juntao I was going through 10-Q reports and found that it's mostly divided into 2 parts with 1st part containing the financial information and second part containing things that...

Just to be clear, the blogs should be hosted on a personal website instead of platforms like medium?

Hi, I was following this [guide](https://github.com/kubeedge/ianvs/blob/main/examples/robot/lifelong_learning_bench/semantic-segmentation/README.md) but am unable to get past the download dataset command. Can you please provide a link from where I can get the dataset mentioned...