studio: improve training progress messages for large datasets

Add granular status updates through the full preprocessing pipeline:
- "Downloading dataset: Open-Orca/OpenOrca..." before HF download
- "Downloaded Open-Orca/OpenOrca (4,233,923 rows)" after download
- "Formatting dataset (4,233,923 rows)..." before format step
- "Applying chat template to chatml_conversations (4,233,923 rows)..."
- "Dataset ready (4,233,923 samples, chatml_conversations format)"

Shows detected format name and row counts at each stage so users
can see progress through large dataset preprocessing instead of
a static "Loading and formatting dataset..." for minutes.

studio/backend/core/training/trainer.py

@@ -2345,6 +2345,9 @@ class UnslothTrainer:
                         status_message = f"Streamed {len(dataset)} rows from HuggingFace"
                     )
                 else:
+                    self._update_progress(
+                        status_message = f"Downloading dataset: {dataset_source}..."
+                    )
                     dataset = load_dataset(**load_kwargs)
 
                 # Check if stopped during dataset loading
@@ -2354,7 +2357,7 @@ class UnslothTrainer:
 
                 n_rows = len(dataset) if hasattr(dataset, "__len__") else 0
                 self._update_progress(
-                    status_message = f"Loaded dataset from HuggingFace: {dataset_source} ({n_rows:,} rows)"
+                    status_message = f"Downloaded {dataset_source} ({n_rows:,} rows)"
                 )
                 logger.info(
                     f"Loaded dataset from Hugging Face: {dataset_source} ({n_rows:,} rows)\n"
@@ -2482,10 +2485,13 @@ class UnslothTrainer:
                 self._update_progress(error = error_msg)
                 return None
 
+            detected = dataset_info.get("detected_format", "unknown")
+            final_ds = dataset_info.get("dataset")
+            final_n = len(final_ds) if hasattr(final_ds, "__len__") else "?"
             self._update_progress(
-                status_message = f"Dataset formatted and ready for training"
+                status_message = f"Dataset ready ({final_n:,} samples, {detected} format)"
             )
-            logger.info(f"Dataset formatted successfully\n")
+            logger.info(f"Dataset formatted successfully ({final_n} samples, {detected})\n")
 
             # ========== THEN SPLIT ==========
             if has_separate_eval_source and eval_dataset is not None:

studio/backend/utils/datasets/dataset_utils.py

@@ -1109,9 +1109,10 @@ def format_and_template_dataset(
         )
 
         # Step 2: Apply chat template
+        detected = dataset_info.get("detected_format", "unknown")
         if progress_callback and n_rows:
             progress_callback(
-                status_message = f"Applying chat template ({n_rows:,} rows)..."
+                status_message = f"Applying chat template to {detected} ({n_rows:,} rows)..."
             )
         # Gemma emits a leading <bos> that must be stripped for text-only chatml/sharegpt.
         is_alpaca = format_type == "alpaca" or (