Properly support batched/non-batched with vllm/llama.cpp #77

Closed

njhill wants to merge 10 commits intoinstructlab:mainfrom

njhill:remove_iterblock

Commits on Jul 3, 2024

Remove the iterBlock and use openai's 'n' parameter instead

npalaska
authored and
njhill
committed
some debug

npalaska
authored and
njhill
committed
fix zipping of samples and outputs

npalaska
authored and
njhill
committed
some refactoring

npalaska
authored and
njhill
committed
fix the num_samples location

npalaska
authored and
njhill
committed
update generate_data API calls

npalaska
authored and
njhill
committed
change back the defaults in test scripts

npalaska
authored and
njhill
committed
fix SimpleFlows

npalaska
authored and
njhill
committed
Properly support batched/non-batched with vllm/llama.cpp
njhill
committed
Address review comments
njhill
committed