Hello everyone,
I would like to solve the following problem with the Google Cloud, but I keep getting only errors. The following scenario: I need a machine learning model that assigns prompts made to a chatbot to a defined category. To do this, I have manually annotated a clean data set with 1300 entries in a table (prompt, category).
I have already tried to create a tabular category set and text dataset for this using the VertexAI dataset menus. If I use the tabular variant, I get the error: Column specifications cannot be retrieved. This is despite the fact that the columns in the .csv are separated by commas.
When creating the text dataset by converting to JSON, strangely enough, no columns and values are recognized either. The JSON looks like this:
{
"Prompt": "text promp1",
"category": "manually annotated category for 1"
},
{
"Prompt": "text promp1",
"category": "manually annotated category for 2"
},
I am new to this and would appreciate any feedback. Maybe I am still doing something fundamentally wrong.
Many thanks in advance for your help
LinkedIn
Twitter
