Update README.md

README.md

@@ -32,11 +32,9 @@ pipeline_tag: text-generation
 
 This is based on the Quiet Star Reasoning Project : which was abandoned earlier in the year :)
 
-Current update : UNDER TEST ! / Self Extend still to be applied : this is just an Early Release of the StaR model ! cleaned up tensors but !
-problem with cos cache ?? --- Will fix it tomorow or this week here : Im workingonit!
-Also cannot get to do the trainingbecause it need large memory and the A100 colab will not seem to work ! ( accelerate issue )
-(I cannot make it a gguf ?HOW!) - Unless maybe i hack the transformers library maybe ?(reframe the model as themistral and replace the existing file ( thats how they had doen it in the past , perhaps i will have to stay as a remte code model ?))
-
+Current update : UNDER TEST ! Currently loading in unsloth but there still is the cos-cache issue so the model will load as mistral :
+to load as talking heads you will need to copy modelling.py/configuration.py to the mistral folder of the transformers library and compile the latest then it loads fine but as remote code it has this issue !
+i will FIX IT !!!~ 
 # Introduction :
 
 ## STAR REASONERS !