Update app.py

app.py

@@ -29,14 +29,17 @@ def upload_visibility(radio):
     else:
         return gr.UploadButton(visible=bool(0))
 
-def get_dataset(dataset_predefined, dataset_upload, data_field):
+#from datasets import load_dataset
+
+def get_predefined_dataset(dataset_name):
+    dataset = load_dataset(dataset_name, split = "train")
+    return dataset
 
-    if dataset_predefined.visible:
-        print('')
-    else:
-        print('')
-    return 0
 
+def get_uploaded_dataset():
+    with open(file.name, 'r') as f:
+        content = f.read()
+    return content[0:100]
 
 
 
@@ -80,7 +83,7 @@ def main():
         ##### Title Block #####
         gr.Markdown("# Instruction Tuning with Unsloth")
     
-        ##### Model Inputs #####
+        ##### Initial Model Inputs #####
         gr.Markdown("### Model Inputs")
         
         # Select Model
@@ -98,18 +101,31 @@ def main():
         dataset_predefined = gr.Textbox(label="Hugging Face Hub Dataset", 
                                         value='yahma/alpaca-cleaned', 
                                         visible=True)
-        dataset_upload = gr.UploadButton(label="Upload Dataset (csv, jsonl, or txt)", 
+        dataset_predefined_load = gr.Button("Upload Dataset")
+        dataset_uploaded_load = gr.UploadButton(label="Upload Dataset (csv, jsonl, or txt)", 
                                          file_types=[".csv",".jsonl", ".txt"], 
                                          visible=False)
         data_field = gr.Textbox(label="Dataset Training Field",
                                 value=conf['model']['general']["dataset_text_field"])
+        data_snippet = gr.Markdown()
         dataset_choice.change(textbox_visibility, 
                               dataset_choice, 
                               dataset_predefined)
         dataset_choice.change(upload_visibility, 
                               dataset_choice, 
                               dataset_upload)
+        
+        # Dataset button
+        dataset_predefined_load.click(fn=get_predefined_dataset
+                                  inputs=dataset_predefined_load,
+                                  outputs=data_snippet)
+
+        dataset_uploaded_load.click(fn=get_uploaded_dataset,
+                                 inputs=dataset_uploaded_load,
+                                 outputs=data_snippet)
+
 
+        ##### Model Parameter Inputs #####
         gr.Markdown("### Model Parameter Selection")
         # Parameters
         max_seq_length = gr.Textbox(label="Maximum sequence length", 
@@ -138,7 +154,7 @@ def main():
         
         ##### Execution #####
     
-        # Setup button
+        # Setup buttons
         tune_btn = gr.Button("Start Fine Tuning")
         gr.Markdown("### Model Progress")
         # Text output (for now)
@@ -148,7 +164,7 @@ def main():
         # Data retrieval
         
         
-        # Execute button
+        # Execute buttons
         tune_btn.click(fn=train, 
                        inputs=[model_name, 
                                inject_prompt,