[Examples] Add SFT disclaimers (#1640)

## Purpose ##
* Add disclaimers about SFT, which is more effective in other training
libraries
* Related to #1638

Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>

Author: kylesayrs

examples/quantization_2of4_sparse_w4a16/llama7b_sparse_w4a16.py

@@ -1,3 +1,7 @@
+# NOTE: Fine tuning can require more steps than is shown in the example
+# See the Axolotl integration blog post for best fine tuning practices
+# https://developers.redhat.com/articles/2025/06/17/axolotl-meets-llm-compressor-fast-sparse-open
+
 from pathlib import Path
 
 import torch
@@ -74,6 +78,7 @@
 )
 
 # Sparse finetune
+# This step can be supplanted by fine tuning via integrated FT libraries such as Axolotl
 train(
     model=(output_path / "sparsity_stage"),
     **oneshot_kwargs,

examples/trl_mixin/ex_trl_constant.py

@@ -1,3 +1,7 @@
+# NOTE: Fine tuning can require more steps than is shown in the example
+# See the Axolotl integration blog post for best fine tuning practices
+# https://developers.redhat.com/articles/2025/06/17/axolotl-meets-llm-compressor-fast-sparse-open
+
 from datasets import load_dataset
 from sft_trainer import SFTTrainer
 from transformers import AutoModelForCausalLM, AutoTokenizer
@@ -46,6 +50,7 @@ def formatting_prompts_func(example):
 )
 model_args = ModelArguments(model=model)
 
+# This step can be supplanted by fine tuning via integrated FT libraries such as Axolotl
 trainer = SFTTrainer(
     model=model,
     processing_class=tokenizer,

examples/trl_mixin/ex_trl_distillation.py

@@ -1,3 +1,7 @@
+# NOTE: Fine tuning can require more steps than is shown in the example
+# See the Axolotl integration blog post for best fine tuning practices
+# https://developers.redhat.com/articles/2025/06/17/axolotl-meets-llm-compressor-fast-sparse-open
+
 from sft_trainer import SFTTrainer
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
@@ -60,6 +64,7 @@
 )
 model_args = ModelArguments(model=model, distill_teacher=teacher)
 
+# This step can be supplanted by fine tuning via integrated FT libraries such as Axolotl
 trainer = SFTTrainer(
     model=model,
     teacher=teacher,