Use wikitext-2-raw-v1 in the examples (#3333)

alexsu52 · web-flow · commit be26567647e7 · 2025-03-11T12:59:54.000+04:00
### Changes Use wikitext-2-raw-v1 in the examples ### Reason for changes Minimize number downloads. ### Related tickets N/A ### Tests https://github.com/openvinotoolkit/nncf/actions/runs/13715504451
diff --git a/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py b/examples/llm_compression/openvino/tiny_llama_find_hyperparams/main.py
@@ -261,7 +261,7 @@ def main():
         load_in_8bit=False,
     )
     tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
-    dataset = load_dataset("wikitext", "wikitext-2-v1", split="train[:1000]")  # <YOUR_DATASET>
+    dataset = load_dataset("wikitext", "wikitext-2-raw-v1", split="train[:1000]")  # <YOUR_DATASET>
     dataset = dataset.filter(lambda example: len(example["text"]) > 128)
     transform_func = partial(tiny_llama_transform_func, tokenizer=tokenizer, ov_model=model.model)
 

Original file line number	Diff line number	Diff line change
`@@ -261,7 +261,7 @@ def main():`
`261`	`261`	`load_in_8bit=False,`
`262`	`262`	`)`
`263`	`263`	`tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)`
`264`		`- dataset = load_dataset("wikitext", "wikitext-2-v1", split="train[:1000]") # <YOUR_DATASET>`
	`264`	`+ dataset = load_dataset("wikitext", "wikitext-2-raw-v1", split="train[:1000]") # <YOUR_DATASET>`
`265`	`265`	`dataset = dataset.filter(lambda example: len(example["text"]) > 128)`
`266`	`266`	`transform_func = partial(tiny_llama_transform_func, tokenizer=tokenizer, ov_model=model.model)`
`267`	`267`