Update README.md

README.md

@@ -11,23 +11,15 @@ Edit this `README.md` markdown file to author your organization card 🔥
 
 *Chatgpt generated summary*
 
-Workflow Steps
-Database Generation
-
-A space named 'database-generation' accepts a .jsonl file of inputs and produces a file of the ChatGPT generations.
-This output is stored in a separate Hugging Face 'dataset' repo, circumventing spaces' lack of persistent memory.
-A user interface opens at the end for review of generations.
-This space runs on a CPU and thus is free.
-Model Training
-
-In this space, you select the dataset for training which it fetches from the HF 'dataset' repo.
-Post-training, it stores the new fine-tuned model in a HF 'models' repo.
-Another user interface opens for reviewing the fine-tuned generations.
-The space has a 15-minute idle timeout.
-Model Inference
-
-You select the fine-tuned model for use in the readme of this space.
-If no model is specified, it uses the base (non-fine-tuned) model.
-Important Notes
-Any change, even to the readme, triggers a full restart of the space.
-The script, upon completion, will re-run in a continuous cycle.
+Workflow Summary
+
+Database Generation:
+
+-  jsonl file (inputs) -> ChatGPT generations -> HF 'dataset' repo
+Model Training:
+
+- Selected dataset from HF 'dataset' repo -> Fine-tuned model -> HF 'models' repo
+
+Model Inference:
+
+- Selected fine-tuned model (or base model if none selected) -> Inference results