Update README.md

README.md

@@ -21,21 +21,26 @@ widget:
       ### Instruction:
       Tell me all you know about the Earth.
       
+      ### Response:
+  - text: |-
+      ### Instruction:
+      How to become more healthy?
+      
       ### Response:
 inference:
   parameters:
     max_new_tokens: 32
-    repetition_penalty: 1.15
+    repetition_penalty: 1.18
     do_sample: true
-    temperature: 0.5
-    top_p: 0.15
+    temperature: 0.7
+    top_p: 0.1
+    top_k: 40
 ---
 
 # Locutusque's TinyMistral-248M trained on the Evol Instruct dataset using AutoTrain
 
 - Base model: [Locutusque/TinyMistral-248M](https://huggingface.co/Locutusque/TinyMistral-248M)
 - Dataset: [KnutJaegersberg/WizardLM_evol_instruct_V2_196k_instruct_format](https://huggingface.co/datasets/KnutJaegersberg/WizardLM_evol_instruct_V2_196k_instruct_format)
-- [Training hyperparameters](https://huggingface.co/Felladrin/TinyMistral-248M-Evol-Instruct/blob/85723f768c439307c996328d68fe7df2947237ae/training_params.json)
 - Availability in other ML formats:
   - GGUF: [afrideva/TinyMistral-248M-Evol-Instruct-GGUF](https://huggingface.co/afrideva/TinyMistral-248M-Evol-Instruct-GGUF)