Use correct model (#1810)

SUMMARY:
This test was skipped and when run, was giving the can't copy out of
meta tensor error. This is because the
[model](https://huggingface.co/Xenova/llama2.c-stories110M/discussions/4)
is missing lm_head and embed_token layers and therefore only loaded to
meta device. Switching to our model to avoid this issue.


TEST PLAN:
Tested locally and passed.

Signed-off-by: shanjiaz <[email protected]>
Co-authored-by: Brian Dellabetta <[email protected]>

Author: shanjiaz

tests/llmcompressor/recipe/test_recipe_parsing.py

@@ -17,7 +17,7 @@ def setup_model_and_config(tmp_path):
     Loads a test model and returns common arguments used in oneshot runs.
     """
     model = AutoModelForCausalLM.from_pretrained(
-        "Xenova/llama2.c-stories110M",
+        "nm-testing/llama2.c-stories15M",
         torch_dtype="auto",
     )