wassname/Open-Assistant
1
0
Fork 0
mirror of https://github.com/wassname/Open-Assistant.git synced 2026-08-20 12:00:32 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix typos (#1143)

Browse Source 
Found via `codespell -S .mypy_cache,yarn.lock,*.json,*.ipynb -L
rouge,nam,vie`
This commit is contained in:
Kian-Meng Ang
2023-02-05 20:18:03 +01:00
committed by GitHub
parent 364a4f5aa3
commit 1e321a6fca
32 changed files with 50 additions and 50 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/data-collection/twitter/README.md
+2 -2
View File
@@ -57,7 +57,7 @@ conversation, or at least as a prompt with replies.
guarantee of the quality of the tweets.
- The tweet quality is the other major issue. We can get conversations through
the currently made scripts, but they most likely don't match a useful
instruction -> fulfilment. We are trying to filter the tweets through various
instruction -> fulfillment. We are trying to filter the tweets through various
means such as matching useful hashtags, or by using cosine similarity against
known instructions.
- The modern Twitter API has conversation_id as a field which can be a way to
@@ -68,7 +68,7 @@ conversation, or at least as a prompt with replies.
## TODO
- Write scripts to filter existing conversations into useful instructions ->
fulfilment with hashtags or cosine similarity.
fulfillment with hashtags or cosine similarity.
- Train model to detect if text is a suitable instruction. This could then be
run through the conversations (or full tweet dump) to simplify the process.
Related to issue #143.