🦖 𝗟𝗲𝗮𝗿𝗻 about 𝗟𝗟𝗠𝘀, 𝗟𝗟𝗠𝗢𝗽𝘀, and 𝘃𝗲𝗰𝘁𝗼𝗿 𝗗𝗕𝘀 for free by designing, training, and deploying a real-time financial advisor LLM system ~ 𝘴𝘰𝘶𝘳𝘤𝘦 𝘤𝘰𝘥𝘦 + 𝘷𝘪𝘥𝘦𝘰 & 𝘳𝘦𝘢𝘥𝘪𝘯𝘨 𝘮𝘢𝘵𝘦𝘳𝘪𝘢𝘭𝘴
MIT License
3k
stars
462
forks
source link
[OfflinePipeline-2] Adapt train/inference to new Q&A dataset format #6
updated training_data.json,testing_data.json dataset files to follow the new format:
-- user_description : short user description (e.g I'm a graduate student)
-- alpaca_context : injected context from alpaca_news (e.g Stock market is on the rise, bull run might come next year)
-- question : user question (e.g Is it a good time to look into retail investments?)
-- response : desired FinanceBot response
Added PrompterTemplate class to handle prompt formatting for both inference/training jobs.
CHANGELOG:
updated
training_data.json
,testing_data.json
dataset files to follow the new format: --user_description
: short user description (e.g I'm a graduate student) --alpaca_context
: injected context from alpaca_news (e.g Stock market is on the rise, bull run might come next year) --question
: user question (e.g Is it a good time to look into retail investments?) --response
: desired FinanceBot responseAdded
PrompterTemplate
class to handle prompt formatting for both inference/training jobs.Renamed old
FinQA
toQA
Added dataset samples to respect
1/5
test/train split (current sizes 25/125)Adapted comet.ml LLM logging to log: -- Question/Answer -- Prompt Template -- Prompt Template Arguments
Ran test of comet.ml LLM prompts logging
Ran test of comet.ml model logging