pipelines go to gravenhollow now.

qlora_pdf_pipeline.py

@@ -2,7 +2,7 @@
 """
 QLoRA Fine-Tuning Pipeline – Kubeflow Pipelines SDK
 
-Fetches PDFs from a Quobjects S3 bucket, extracts instruction-tuning
+Fetches PDFs from a RustFS S3 bucket, extracts instruction-tuning
 data, trains a QLoRA adapter on the Llama 3.1 70B base model using
 the Strix Halo's 128 GB unified memory, evaluates it, and pushes the
 adapter weights to a Gitea repository.
@@ -23,7 +23,7 @@ from typing import NamedTuple
 
 
 # ──────────────────────────────────────────────────────────────
-# 1. Fetch PDFs from Quobjects S3
+# 1. Fetch PDFs from RustFS S3
 # ──────────────────────────────────────────────────────────────
 @dsl.component(
     base_image="python:3.13-slim",
@@ -36,7 +36,7 @@ def fetch_pdfs_from_s3(
     aws_access_key_id: str,
     aws_secret_access_key: str,
 ) -> NamedTuple("PDFOutput", [("pdf_dir", str), ("num_files", int)]):
-    """Download all PDFs from a Quobjects S3 bucket."""
+    """Download all PDFs from an S3 bucket."""
     import os
     import boto3
     from botocore.client import Config
@@ -571,7 +571,7 @@ def log_training_metrics(
                 "learning_rate": learning_rate,
                 "num_epochs": num_epochs,
                 "num_pdfs": num_pdfs,
-                "data_source": "quobjects/training-data",
+                "data_source": "rustfs/training-data",
             }
         )
         mlflow.log_metrics(
@@ -591,13 +591,13 @@ def log_training_metrics(
 @dsl.pipeline(
     name="QLoRA PDF Fine-Tuning",
     description=(
-        "Fine-tune Llama 3.1 70B via QLoRA on PDFs from the Quobjects "
+        "Fine-tune Llama 3.1 70B via QLoRA on PDFs from the RustFS "
         "training-data bucket. Pushes the adapter to Gitea and logs "
         "metrics to MLflow."
     ),
 )
 def qlora_pdf_pipeline(
-    # ── S3 / Quobjects ──
+    # ── S3 / RustFS ──
     s3_endpoint: str = "https://gravenhollow.lab.daviestechlabs.io:30292",
     s3_bucket: str = "training-data",
     s3_prefix: str = "",