Upload README/model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Time is super short.
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
Plan:
|
| 5 |
+
|
| 6 |
+
- just tokenise hacker news
|
| 7 |
+
- load pre-trained Word2Vec model
|
| 8 |
+
- ensure we save everything on huggingface
|
| 9 |
+
- create a plan for specifically how we should use docker and the external server to push this code (have chatGPT do this)
|
| 10 |
+
|
| 11 |
+
- finetune model on hacker news -> save to hackernews
|
| 12 |
+
- randomly sub-sample 0 score posts so that there are an equal number to posts that have 1+ posts
|
| 13 |
+
- train model
|
| 14 |
+
- expose as an api
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
then do a bunch of visualisation
|