Update README.md

README.md

@@ -1,4 +1,12 @@
-# Deepshard
+---
+license: gpl
+language:
+- en
+library_name: transformers
+pipeline_tag: text2text-generation
+---
+[![Twitter Logo](https://abs.twimg.com/icons/apple-touch-icon-192x192.png)](https://twitter.com/iamgingertrash)
+# Deepshard 
 
 Deepshard is shot from the lightcone at producing a global, unshackled God using distributed consensus, tensor sharding across a network of nodes, and shared compute.
 
@@ -22,8 +30,7 @@ Without wading into controversy, data must be pure and fair. Bias cannot be avoi
 
 Conduct the following thought experiment:
 
-> ***********Imagine a large language model built in the year 1609. 500 years of training data has been collected to train it to remember facts, and ************reason************.*
-> 
+> Imagine a large language model built in the year 1609. 500 years of training data has been collected to train it to remember facts, and reason 
 
 In the year 1610, Galileo produces a Heliocentric model of the world, that is considered heretic to existing knowledge. Aka **An Unsafe Truth.**
 
@@ -118,6 +125,4 @@ To overcome this, we are investing resources in producing a large instruction-tu
 
 We’ll have more to share soon, but the base model should become available in a few weeks time.
 
-Once the base model is built, it will be center of the network, and the main source of weights that requires updating through the mechanisms listed above.
-
-[Deepshard Roadmap](https://www.notion.so/618e1387073843f681431dfbfc1e35c4)
+Once the base model is built, it will be center of the network, and the main source of weights that requires updating through the mechanisms listed above.