DuaSPHysics Cell Division Code Structure Try to straight up the cell division part of DuaSPHysics's tricky code. A structure map is shown. Class inheriting JSphGpuSingle::Run ConfigCellDivision(); void J 2023-08-17 #CUDA
Build and train an opensource chatGPT locally This is a conclusion of the last 5 blog, showing the key process of building and optimizing a chat bot backended by open source large language model. 2023-04-18 #deep learning #llama
Possible applications on Car industry of the language model A collection of possible applications of language models, including car design for now. 2023-04-14 #deep learning #llama
scale up and benchmark the language model The process to further improve the current alpaca model by scaling the model to more parameters. A detailed cross-benchmark among 7B, 13B, 30B, chatGPT is followed. 2023-04-12 #deep learning #llama
Improve the chatbot The process to further improve the current alpaca model with larger datasets. 2023-04-02 #deep learning
Chinese chatbot summary A collection of recent models variants from llama/alpaca that can be considered later, including the Chinese chatbot. 2023-04-01 #deep learning
Fine tuning chatbot This is a process of optimizing the chat bot that I just build in last blog. We first make the model more like chatGPT by few shot learning, then we fine tune the model follow Stanford's 2023-03-30 #deep learning
build a chatbot backended by Meta LLaMA A castrated version of chatGPT is built backended by LLaMA, Meta's leaked large language model. Vue.js and flask is used for front-end and backend, and axios is used for HTTP requests. 2023-03-22 #deep learning
WNPS user's guide This is a brief user’s guide of the Wind Noise Prediction Software (WNPS). This project aimed to predict the wind noise propagation paths of the holes & gaps and provide guidance for 2023-02-24 #manual
test GPU server multiGPU performance Following last blog, a benchmark is performed on 8 Nvidia A800 GPU server. The P2P bandwidth and multi GPU performance of N-body problem are tested. 2023-02-01 #CUDA