-
Specialized full-fledged databases for unstructured data.
-
How to imporove retrieval performance?
- Embedding Model Selection
- Do i have the right embedding model for my data?
- Do my embeddings capture BOTH my documents and queries?
- Document Storage Strategy
- Should i store the whole document as one? Or split up into chunks?
- Embedding Model Selection
-
Other Resources:
- dfd
- Do you actually need a vector database? by Ethan Rosenthal
- https://learning.edx.org/course/course-v1:Databricks+LLM101x+2T2023/block-v1:Databricks+LLM101x+2T2023+type@sequential+block@a7512e678c8a4fd98ae3635977fb55bb/block-v1:Databricks+LLM101x+2T2023+type@vertical+block@b475e8c705f448818785a4d5dc2952e4
-
Typical Flavors or fine tuned versions:
- Base - I know what word comes next
- Chat - I know how to engage in conversation
- Instruct - I know how to respond to instructions
-
Multiple Sizes:
- Small
- Base
- Large
-
Which one to use? Need to balance below three:
- Accuracy
- Speed
- Task Specific Performance in a3.5
- Fine tuning using Dollyv2
- Need to look at task specific evaluation.
Natural language answer vs Natural language groud truch and compare them for semantic equivalence.