Upload folder using huggingface_hub

ui.py

@@ -14,7 +14,7 @@ def build_interface(engine: FunctionGemmaEngine) -> gr.Blocks:
                 
                 with gr.Row():
                     with gr.Column(scale=1):
-                        gr.Markdown("**Step 1: Define Functions**\n\nEdit the JSON schema below to define the tools the model should learn.")
+                        gr.Markdown("**Step 1: Define Functions**<br>Edit the JSON schema below to define the tools the model should learn.")
                         tools_editor = gr.Code(
                             value=engine.get_tools_json(), 
                             language="json", 
@@ -25,8 +25,8 @@ def build_interface(engine: FunctionGemmaEngine) -> gr.Blocks:
                         tools_status = gr.Markdown("")
 
                     with gr.Column(scale=1):
-                        gr.Markdown("**Step 2: Upload Data (Optional)**\n\nUpload a CSV file to replace the default dataset ([bebechien/SimpleToolCalling](https://huggingface.co/datasets/bebechien/SimpleToolCalling)).")
-                        gr.Markdown("**Example CSV Row:**<br>Format: `[User Prompt, Tool Name, Tool Args JSON]`\n```csv\n\"What is the weather in London?\", \"get_weather\", \"{\"\"location\"\": \"\"London, UK\"\"}\"\n```")
+                        gr.Markdown("**Step 2: Upload Data (Optional)**<br>To train on your own data, upload a CSV file to replace the [default dataset](https://huggingface.co/datasets/bebechien/SimpleToolCalling).")
+                        gr.Markdown("**Example CSV Row:** No header required.<br>Format: `[User Prompt, Tool Name, Tool Args JSON]`\n```csv\n\"What is the weather in London?\", \"get_weather\", \"{\"\"location\"\": \"\"London, UK\"\"}\"\n```")
                         import_file = gr.File(
                             label="Upload Dataset (.csv)", 
                             file_types=[".csv"], 
@@ -57,7 +57,7 @@ def build_interface(engine: FunctionGemmaEngine) -> gr.Blocks:
                         param_shuffle = gr.Checkbox(
                             value=True,
                             label="Shuffle Data",
-                            info="Randomize before split"
+                            info="Randomize before split\nWhen unchecking this for your own custom datasets, ensure your source data is pre-mixed. If the distribution is unknown or sorted, you should check this to ensure the model learns a balanced representation of all tools during training."
                         )
 
                 with gr.Row():