readme.md

Performance Evaluation Llama-7B

1. Llama-7b model performanc is good with theoretical questions and prompts generations.
2. Able to solve fundamental csv problems
3. Stucks in more complex problem requiring more logical reasoning
4. Can provide good results with user database.
5. Performance is good when producing multiple results from same prompt, novelty is maintained.

Performance Evaluation of RedPajama-3b

1. Able to answer Open source type question
2. For Context Aware Questions, sometime answers are irrelevant, sometime not able to reproduce results
3. Very poor performance on multiple answer questions.

Performance Evaluation Bloke-Syntara-7B

1. Good with theoretical and prompts generation questions.
2. Not able to solve simple mathematical problems from dataframe/csv
3. Performance is good when producing multiple results from same prompt, novelty is maintained.
4. Performs on Code generation is also better than llama-7b