Mastering Langchain course by Sharath Raju

https://learning.oreilly.com/course/langchain-masterclass/9781835464427/

https://github.com/PacktPublishing/LangChain-MasterClass---Build-15-OpenAI-and-LLAMA-2-LLM-Apps-using-Python/tree/main

Topics covered:

Streamlit:
For quick web UI prototyping

OpenAI LLM, audio transcription.
HuggingFace hosted models

Google Colab - for using huggingface transformers which download the model locally.

Running Llama 2 locally via CTransformers or via Replicate API

DuckduckGo search

Prompt Templates:
FewShotPromptTemplate, Length based example selector


Splitters:
RecursiveCharacterTextSplitter
CharacterTextSplitter

Embeddings:
SentenceTransformerEmbeddings
OpenAIEmbeddings
similarity_search

Response formatter/output parser:
csv
json

Vector DBs:
Pinecone - hosted
Chroma - local
FAISS(not typically used as vector db, but for similarity search etc)

Memory:
ConversationSummaryMemory
TokenBuffer
WindowBuffer
ConversationBuffer

Agents:
Pandas dataframe agent - to run some python code to analyze csv data

Chains:
Utility Chains:
Load_Summarize_Chain - load and summarize docs
LLMRequestsChain - to fetch a url


Others:
SequentialChain
chain_type=stuff/map_reduce

Building Microservices by Sam Newman

I wrote the bulk of this book throughout the whole of 2020 and in the first half of 2021. I used Visual Studio Code on macOS for most of the writing, although on rare occasions I also used Working Copy on iOS. OmniGraffle was used to create all the diagrams in the book. AsciiDoc was used to format the book and on the whole was excellent, with O’Reilly’s Atlas toolchain doing the magic of making a book appear on the other side.

-------------------------------------------