Dhar Rawal

4k tokens is roughly 3000 words, so llama 2 4k context might work. You won't know until you try

Here is a how-to question from my specific domain with the exact answer and the truncated answer from DSPy: **Question**: How do I upgrade my RadiantOne Platform? **DSPy truncated answer**:...

@detaos that worked! I will now test scenarios where the long form answer is assembled from multiple short answers gathered from different contexts. I am not sure they fit within...

@arnavsinghvi11 - I reverted all the complicated changes in PR #843 so that can be merged and this PR can be closed

@arnavsinghvi11 - I have reverted ALL the complicated changes. Hopefully now this PR is acceptable

@arnavsinghvi11 - BootstrapWithRandomSearch changes are now reverted

Ah! thx for the insight. Now I understand why the current code does not work. Let's say you have an input field without content. In this case, the code will...

So... the correct fix looks like below: ``` for completion in completions: if all((completion.get(key, "") != "") for key in field_names if key not in example): finished_completions.append(completion) continue finished_completions.append( extend_generation(completion,...

@mikeedjones No. Its an input field that can be empty

My proposed fix above is not enough. When there is a legitimate extend generations scenario but there is an empty input field, the code still gets into a loop and...