From 9a1531da90c5b32dadc4c64627a1083d6e938e8e Mon Sep 17 00:00:00 2001
From: Tai An <antai12232931@outlook.com>
Date: Thu, 18 Jun 2026 18:18:42 -0700
Subject: [PATCH] docs(finetuning): remove deprecated torchtune references in
 QLoRA and float8 sections

Follow-up to #4312 (per @andrewor14's review request). torchtune development
wound down in 2025 and the library is no longer actively maintained, so the
remaining torchtune-specific recipes are removed:

- QLoRA: drop the deprecated "Option 1: TorchTune Integration" recipe and
  promote the HuggingFace PEFT integration to Option 1. torchao's NF4 usage is
  still demonstrated directly via the FrozenNF4Linear example above the options.
- Float8: drop the deprecated `tune run full_finetune_distributed` recipe and
  reword to point at the shared float8 training APIs (TorchTitan). The measured
  throughput/accuracy results and the pre-training tutorial pointer are kept.

Signed-off-by: Tai An <antai12232931@outlook.com>
---
 docs/source/eager_tutorials/finetuning.rst | 29 +++-------------------
 1 file changed, 3 insertions(+), 26 deletions(-)

diff --git a/docs/source/eager_tutorials/finetuning.rst b/docs/source/eager_tutorials/finetuning.rst
index 7332715e24..0de283124c 100644
--- a/docs/source/eager_tutorials/finetuning.rst
+++ b/docs/source/eager_tutorials/finetuning.rst
@@ -290,20 +290,7 @@ such as regular INT4 or even newer `MXFP4 or NVFP4 <https://github.com/pytorch/a
 targeting Blackwell GPUs to reap similar memory benefits with
 varying tradeoffs.
 
-Option 1: TorchTune Integration
-===============================
-
-TorchTune incorporates the `NF4Tensor` in its QLoRA fine-tuning
-recipe through their implementation of `LoRALinear <https://github.com/pytorch/torchtune/blob/a6290a5b40758f13bca61c386bc8756a49ef417e/torchtune/modules/peft/lora.py#L19>`__.
-You can also try it out by running the following command,
-or refer to their `QLoRA tutorial <https://docs.pytorch.org/torchtune/stable/tutorials/qlora_finetune.html>`__
-for more details.
-
-.. code::
-
-  tune run lora_finetune_single_device --config llama3_2/3B_qlora_single_device.yaml
-
-Option 2: HuggingFace PEFT Integration
+Option 1: HuggingFace PEFT Integration
 ======================================
 
 `HuggingFace PEFT <https://huggingface.co/docs/peft/main/en/developer_guides/quantization#torchao-pytorch-architecture-optimization>`__
@@ -332,18 +319,8 @@ Float8 Quantized Fine-tuning
 Similar to `pre-training <pretraining.html>`__, we can also
 leverage float8 in fine-tuning for higher training throughput
 with no accuracy degradation and no increase in memory usage.
-Float8 training is integrated into TorchTune's distributed
-full fine-tuning recipe, leveraging the same APIs as our
-integration with TorchTitan. Users can invoke this fine-tuning
-recipe as follows:
-
-.. code::
-
-  tune run --nnodes 1 --nproc_per_node 4 full_finetune_distributed --config llama3_2/3B_full
-    enable_fp8_training=true \
-    fp8_recipe_name=tensorwise \
-    compile=True
-
+Float8 fine-tuning leverages the same float8 training APIs as our
+integration with `TorchTitan <https://github.com/pytorch/torchtitan>`__.
 Initial experiments saw up to 16.5% throughput improvement
 for fine-tuning Llama3.2-3B in float8: