shmsw25/FActScore issues and pull requests

#48 - My FactScore

Pull Request - State: closed - Opened by IdoAmit198 5 months ago

#47 - Maintaining core factscore in a new repository

Issue - State: open - Opened by armingh2000 6 months ago

#46 - Improved version for speed and openai updates

Issue - State: open - Opened by wj210 6 months ago

#45 - Batching for speed.

Issue - State: open - Opened by wj210 6 months ago

#44 - About the enwiki-20230401

Issue - State: open - Opened by Toblame 7 months ago - 5 comments

#43 - loosen overly restrictive dependencies

Pull Request - State: open - Opened by dojoteef 8 months ago

#42 - make factscore use chatgpt for atomic fact generation

Pull Request - State: open - Opened by dojoteef 8 months ago - 1 comment

#41 - #38 Groundwork for migration to new openai >=v1.0.0 python API

Pull Request - State: closed - Opened by GeroVanMi 8 months ago - 2 comments

#40 - sqlite3.DatabaseError: file is not a database error after Installation

Issue - State: closed - Opened by GeroVanMi 8 months ago

#39 - InvalidRequestError: The model `text-davinci-003` has been deprecated

Issue - State: closed - Opened by shrusti-ghela 8 months ago - 2 comments

#38 - Move to ChatCompletion endpoint for call_GPT3

Issue - State: open - Opened by martiansideofthemoon 8 months ago

#37 - AssertionError: `topic` in your data (Jessie Mae Brown Beavers) is likely to be not a valid title in the DB.

Issue - State: open - Opened by clawnotfound 8 months ago - 1 comment

#36 - problem using the factscorer function with LLaMa

Issue - State: open - Opened by samuel-LP 9 months ago - 1 comment

#35 - Some cases couldn't be found in the default db

Issue - State: open - Opened by Loose-Gu 10 months ago

#34 - Any way to run without an OpenAI key?

Issue - State: open - Opened by mohummedalee 10 months ago - 3 comments

#33 - Issue running get_score() on M1 Mac due to Torch build

Issue - State: closed - Opened by jcblsn 10 months ago - 1 comment

#32 - How to modify the code to calculate factscore between any two paragraphs/sentences without retrieval steps?

Issue - State: open - Opened by laiviet 11 months ago - 1 comment

#31 - Generation details for models under data/labeled

Issue - State: open - Opened by caiqizh 11 months ago - 1 comment

#30 - The chatgpt-based AtomicFactGenerator of FactScore sometimes abstain from answering

Issue - State: closed - Opened by HillZhang1999 11 months ago - 4 comments

#29 - Regarding human evaluation

Issue - State: closed - Opened by pat-jj 11 months ago - 2 comments

#28 - FileNotFoundError (demons.json) for custom knowledge base

Issue - State: open - Opened by MathiasKraus 11 months ago - 2 comments

#27 - Typo here

Issue - State: closed - Opened by CiaoHe 11 months ago - 1 comment

#26 - Scoring existing generated text

Issue - State: closed - Opened by sofi444 about 1 year ago - 1 comment

#25 - Having issues on using LLAMA-7B model

Issue - State: open - Opened by mungg about 1 year ago - 2 comments

#24 - Small changes to ensure custom corpus addition works

Pull Request - State: closed - Opened by martiansideofthemoon about 1 year ago

#23 - FActScore can be slow to run

Issue - State: open - Opened by martiansideofthemoon about 1 year ago

#22 - Using unlabeled data to generate atomic facts and retrieving evidence

Issue - State: closed - Opened by rubaha96 about 1 year ago - 1 comment

#21 - add length penalty

Pull Request - State: closed - Opened by shmsw25 over 1 year ago - 1 comment

#20 - Moving abstain detection to a separate module.

Pull Request - State: closed - Opened by martiansideofthemoon over 1 year ago

#19 - Move abstain detection outside the atomic fact generator

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago

#18 - Create LICENSE

Pull Request - State: closed - Opened by martiansideofthemoon over 1 year ago

#17 - Don’t understand how true_score and false_score are represented in logits

Issue - State: closed - Opened by rubaha96 over 1 year ago - 1 comment

#16 - License?

Issue - State: closed - Opened by brucewlee over 1 year ago

#15 - fix bugs in retrieval

Pull Request - State: closed - Opened by absolute-nil over 1 year ago - 1 comment

#14 - MAX_LENGTH is not defined in retrieval

Issue - State: closed - Opened by absolute-nil over 1 year ago - 2 comments

#13 - Better ranking between models

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago - 1 comment

#12 - Adding cost estimates for OpenAI API usage

Pull Request - State: closed - Opened by martiansideofthemoon over 1 year ago

#11 - setup log, separating data_dir

Pull Request - State: closed - Opened by shmsw25 over 1 year ago

#10 - Improving logging, show different levels

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago

#9 - Instruct LLAMA integration

Pull Request - State: closed - Opened by martiansideofthemoon over 1 year ago

#8 - added `% of responding` and `num_facts_per_response`

Pull Request - State: closed - Opened by shmsw25 over 1 year ago - 1 comment

#7 - Return respond rate and number of facts together with FActScore

Issue - State: closed - Opened by shmsw25 over 1 year ago - 1 comment

#6 - improve processing of Perplexity AI outputs

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago - 2 comments

#5 - Create HF weight diffs for LLAMA model and add it to the README

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago

#4 - Add citation / bibtex to README

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago - 1 comment

#3 - estimated cost for API before running API

Issue - State: closed - Opened by martiansideofthemoon over 1 year ago - 1 comment

#2 - small updates

Pull Request - State: closed - Opened by shmsw25 over 1 year ago

#1 - new data format + atomic fact gen integrated

Pull Request - State: closed - Opened by martiansideofthemoon over 1 year ago

GitHub / shmsw25/FActScore issues and pull requests