Tags from openai-python

v0.10.4: Search v2 functionality (#33)

2021-09-09T17:46:36Z

Search v2 functionality (openai#33)

* Add support for search_indices (openai#64)

* Add support for search_indices

* Updated with Schnurr's comments

* Add version to search (openai#65)

* Make search query required (openai#67)

v0.10.3: Boris/examples and cli (#32)

2021-08-31T15:56:03Z

Boris/examples and cli (openai#32)

* Add a codex backtranslation example to improve SQL queries (openai#58)

* Add a codex backtranslation example to improve SQL queries

* Boris update ft example (openai#57)

* update fine-tune example to show the new CLI outputs

* model specifiction for search (openai#60)

* Catch chunked encoding errors and retry (openai#63)

* Add batch suggestion logic to prepare_data for fine_tunes and custom Q&A answers logic (openai#62)

* Add batch suggestion logic to prepare_data for fine_tunes; add an example of how to create a rudimentary answers endpoint with a custom Q&A model

Co-authored-by: Madeleine Thompson <[email protected]>
Co-authored-by: hallacy <[email protected]>

v0.10.2: Updates to prepare_data function (#29)

2021-07-29T19:10:48Z

Updates to prepare_data function (openai#29)

* update documentation links to point to the website

* Fix encoding

* Add rough time estimator based on historical stats

* Fix train_test split naming logic; add quiet mode for running inside scripts

* Add a finetuning step by step example for a classification use case.

* add classification params if train and valid set; add length_validator

v0.10.1

2021-07-14T21:19:59Z

v0.10.0: renaming snapshot to model (#27)

2021-07-14T20:05:12Z

renaming snapshot to model (openai#27)

* rename snapshot to model in sdk (openai#48)

* updating version

v0.9.4: minor fixes to tools prepare_data validators (#47) (#26)

2021-07-12T20:46:39Z

minor fixes to tools prepare_data validators (openai#47) (openai#26)

* ensure that only a single whitespace is prepended. Ensure the message regarding the prompt separator is displayed only if a prompt separator exists.

* change pandas contains to not use regex, which can trip if the common_suffix is actually a regex

Co-authored-by: Boris Power <[email protected]>

v0.9.3: Cli fixes and improvements (#25)

2021-06-30T19:40:45Z

Cli fixes and improvements (openai#25)

* Revamp cli args (openai#45)

* Rachel/follow (openai#46)

* Add fine_tunes.follow. Add better error handling for disconnected streams

* return early

* fix an oops

* lint

* Nicer strings

* ensure end token is not applied to classification (openai#44)

* ensure end token is not applied to classification

* black

Co-authored-by: Boris Power <[email protected]>

v0.9.2

2021-06-30T04:27:09Z

v0.9.1: bugfix

2021-06-30T04:08:50Z

bugfix

* ensure that pandas empty values are read as empty string, rather than a float

Co-authored-by: Boris Power <[email protected]>

v0.9.0: Lots of CLI changes (#22)

2021-06-29T21:55:00Z

Lots of CLI changes (openai#22)

* Add CLI option to download files (openai#34)

* Option to check if file has been uploaded in the past before uploading (openai#33)

The check is done based on filename, file purpose and file size

* Add fine-tuning hparams directly into the fine-tunes CLI (openai#35)

* update fine_tunes cli use_packing argument (openai#38)

* A file verification and remediation tool.

It applies the following validations:
- prints the number of examples, and warns if it's lower than 100
- ensures prompt and completion columns are present
- optionally removes any additional columns
- ensures all completions are non-empty
- infers which type of fine-tuning the data is most likely in (classification, conditional generation and open-ended generation)
- optionally removes duplicate rows
- infers the existence of a common suffix, and if there is none, suggests one for classification and conditional generation
- optionally prepends a space to each completion, to make tokenization better
- optionally splits into training and validation set for the classification use case
- optionally ensures there's an ending string for all completions
- optionally lowercases completions or prompts if more than a 1/3 of alphanumeric characters are upper case

It interactively asks the user to accept or reject recommendations. If the user is happy, then it saves the modified output file as a jsonl, which is ready for being used in fine-tuning with the printed command.

* Completion: remove from kwargs before passing to EngineAPI (openai#37)

* Version bump before pushing to external

Co-authored-by: Todor Markov <[email protected]>
Co-authored-by: Boris Power <[email protected]>
Co-authored-by: Dave Cummings <[email protected]>