Update README.md

README.md

@@ -12,9 +12,9 @@ license: apache-2.0
 
 # Pre Trained With
 
-- Nostr notes: This makes the model know more about bitcoin and other topics discussed on Nostr. 
 - Health related topics
 - Faith related topics
+- Nostr notes: This makes the model know more about bitcoin and other topics discussed on Nostr. 
 - Nutrition related topics
 - Medicinal herbs
 
@@ -27,22 +27,20 @@ license: apache-2.0
 
 # Uses
 
-You can read more about the "LLM curation" via fine tuning with Nostr notes:
-[here](https://habla.news/a/naddr1qvzqqqr4gupzp8lvwt2hnw42wu40nec7vw949ys4wgdvums0svs8yhktl8mhlpd3qqgx6cnjvfhhzet20pkhqdn2wenkvu6gy4y) and 
-[here](https://habla.news/a/naddr1qvzqqqr4gupzp8lvwt2hnw42wu40nec7vw949ys4wgdvums0svs8yhktl8mhlpd3qqxnzde3xumrswfjx56rjwf4kkqhsx).
-
 Compared to other models this may know more about Nostr, Bitcoin and healthy living. 
 Closer aligment to opinions of people on Nostr. It may have alternative ideas to mainstream because of Nostr and my own curation. 
 So it is basically somewhere in between base llama 3.0 plus Nostr plus my values. 
 
+You can read more about the "LLM curation" via fine tuning with Nostr notes:
+[here](https://habla.news/a/naddr1qvzqqqr4gupzp8lvwt2hnw42wu40nec7vw949ys4wgdvums0svs8yhktl8mhlpd3qqgx6cnjvfhhzet20pkhqdn2wenkvu6gy4y) and 
+[here](https://habla.news/a/naddr1qvzqqqr4gupzp8lvwt2hnw42wu40nec7vw949ys4wgdvums0svs8yhktl8mhlpd3qqxnzde3xumrswfjx56rjwf4kkqhsx).
+
 I am using the model here as a ground truth for Nostr related questions: https://wikifreedia.xyz/based-llm-leaderboard/npub1nlk894teh248w2heuu0x8z6jjg2hyxkwdc8cxgrjtm9lnamlskcsghjm9c
 
 Use repeat penalty of 1.05 or more to avoid repetitions. 
 
 I hope you like it. Let me know about your experience. You can DM me on Nostr.
 
-The number in the filenames (like 21345) means the version. I take the training steps and add those as the version. Each source, book, note adds to the version.
-
 
 # Warning
 
@@ -54,12 +52,17 @@ There is no guarantee that the model will be of any use. It may hallucinate ofte
 
 ## Training Data
 
-The sources mentioned above are converted to TXT files and used as pre training. No PPO, DPO or other method of fine tuning.
+The sources mentioned above are converted to TXT files and used as pre training. A few supervised fine tunings that helped with controlling length of the output.
+
+Sources include like banned videos from youtube. 
 
 ## Training Procedure
 
 LLaMa-Factory is used to train on 2* RTX 3090! fsdp_qlora is the technique. 
 
+The number in the filenames (like 21345) means the version. I take the training steps and add those as the version. Each source, book, note adds to the version.
+
+
 # Contributions
 
 You can tip me on Nostr https://primal.net/p/npub1nlk894teh248w2heuu0x8z6jjg2hyxkwdc8cxgrjtm9lnamlskcsghjm9c