Update README.md

README.md

@@ -7,12 +7,13 @@ license: apache-2.0
 # Model Card for Ostrich
 
 
-Contentious, judgemental, uncensored, can't agree with itself 32% of the time!
-Trained a bit about nostr
-Trained a bit about bitcoin
-Trained a bit in the health domain
+- Contentious, judgemental, uncensored, can't agree with itself 32% of the time!
+- Trained a bit about nostr
+- Trained a bit about bitcoin
+- Trained a bit in the health domain
+
+I am having success with chat template: \<s\> \[INST\] ... \<\/s\> 
 
-I am having success with chat template: <s> [INST] ... </s> 
 It may also work with ChatML format, though I see more repetitions when I use that.
 
 
@@ -42,10 +43,10 @@ The trainer, developer or uploader of this model does not assume any liability.
 
 ### Training Data
 
-Nostr notes, kind=1, longer notes are taken from reputable accounts. 
-Number of notes: 300k
+Nostr related info from web and nostr itself, bitcoin related info.
 
 ### Training Procedure
 
 LLaMa-Factory is used to train on 2x3090! fsdp_qlora is the technique. 
-It took ~170 hours for a dataset of 120MB.
+
+It took ~185 hours for a dataset of 122MB.