Fine Tuning - PostgresML #542
Replies: 2 comments
-
Hello, Trying the fine-tuning example and trying to load the data, I get an error: hasura=# SELECT pgml.load_dataset('kde4', kwargs => '{"lang1": "en", "lang2": "es"}');
ERROR: function pgml.load_dataset(unknown, kwargs => unknown) does not exist
LINE 1: SELECT pgml.load_dataset('kde4', kwargs => '{"lang1": "en", ...
^
HINT: No function matches the given name and argument types. You might need to add explicit type casts.
hasura=# \df pgml.load*
List of functions
Schema | Name | Result data type | Argument data types | Type
--------+--------------+-------------------------------------+--------------------------------------------------+------
pgml | load_dataset | TABLE(table_name text, rows bigint) | source text, "limit" bigint DEFAULT NULL::bigint | func
(1 row) Is there something else that needs to be done to load these sample data? This is for pgml14 and postgresql-14 on Ubuntu 22 |
Beta Was this translation helpful? Give feedback.
-
Fine tuning is the last 1.0 (Python) feature that has not been ported to 2.0 (Rust) yet. If you want to start fine tuning now, you'll need to install 1.0. Some work has been completed to move the core huggingface functionality, (e.g. tokenizers) to Rust, but there is plenty of "task" specific logic to implement for all the various different fine tuning strategies. That part isn't particularly daunting (we could just copy/paste the 1.0 Python into 2.0), but I'm more interested in creating a generic API (implemented in Rust) so that users can define their own fine tuning/text/image/audio/video processing pipelines (beyond pre-built model specific ones already included) via SQL in 2.0. |
Beta Was this translation helpful? Give feedback.
-
Fine Tuning - PostgresML
Train and deploy models to make online predictions using only SQL, with an open source Postgres extension.
https://postgresml.org/user_guides/transformers/fine_tuning/
Beta Was this translation helpful? Give feedback.
All reactions