use llama-3.2-1b in tinygrad test

exo/inference/test_inference_engine.py

@@ -51,6 +51,4 @@ if os.getenv("RUN_TINYGRAD", default="0") == "1":
   import os
   from exo.inference.tinygrad.inference import TinygradDynamicShardInferenceEngine
   tinygrad.helpers.DEBUG.value = int(os.getenv("TINYGRAD_DEBUG", default="0"))
-  asyncio.run(
-    test_inference_engine(TinygradDynamicShardInferenceEngine(NewShardDownloader()), TinygradDynamicShardInferenceEngine(NewShardDownloader()), "llama-3-8b", 32)
-  )
+  asyncio.run(test_inference_engine(TinygradDynamicShardInferenceEngine(NewShardDownloader()), TinygradDynamicShardInferenceEngine(NewShardDownloader()), "llama-3.2-1b", 32))