Update README.md

README.md

@@ -1,20 +1,29 @@
 ---
-title: DockerTestingSpace
-emoji: ✍
+title: FineTunePipelineDocker
+emoji: 🧪
 colorFrom: pink
-colorTo: pink
+colorTo: purple
 sdk: docker
-app_port: 6900
+app_port: 7860
 fullWidth: true
 hf_oauth: true
 tags:
-- argilla
+- transformers
+- finetuning
+- peft
+- gradio
 pinned: true
 license: mit
 ---
-Argilla is a free and open source tool to build and iterate on datasets for AI. It can be deployed on the Hub with a few clicks and Hugging Face OAuth enabled, perfect for running community annotation initiatives!
 
-This is the Argilla Space for:
+This Space runs a lightweight fine-tuning pipeline inside a Docker container.
 
-- Creating your own Argilla Spaces, check the [quickstart guide](http://docs.argilla.io/latest/getting_started/quickstart/) and the [Hugging Face Spaces configuration](http://docs.argilla.io/latest/getting_started/how-to-configure-argilla-on-huggingface/) for more details.
-- Discovering the Argilla UI, sign in with your Hugging Face account!
+What it does:
+- Downloads a dataset from the Hub (default: `karpathy/tiny_shakespeare`)
+- Fine-tunes a small model with LoRA (default base model: `sshleifer/tiny-gpt2`)
+- Saves the trained adapter + tokenizer as artifacts
+- Provides a simple text generation UI for testing
+
+Notes:
+- `app_port` must match the port your container listens on (default is 7860 for Docker Spaces).  
+- OAuth is enabled (`hf_oauth: true`) so you can optionally add “Sign in with HF” and push models on behalf of the user (requires scopes + implementation).