Remove hand authored config, as it now generated by the model builder

.gitignore

@@ -15,6 +15,7 @@ test/csharp/obj/
 /*.csv
 .idea
 cache_dir
+example-models
 *.onnx
 *.onnx.data
 

examples/llama/README.md

@@ -35,7 +35,7 @@ Run the model builder script to export, optimize, and quantize the model. More d
 
 ```bash
 cd examples/phi2
-python -m onnxruntime_genai.models.builder -m meta-llama/Llama-2-7b-chat-hf -e cpu -p int4 -o ./model
+python -m python -m onnxruntime_genai.models.builder -m meta-llama/Llama-2-7b-chat-hf -e cpu -p int4 -o ./example-models/llama2-7b-chat-int4-cpu
 ```
 
 ## Run Llama

examples/llama/llama-loop.py

@@ -1,7 +1,9 @@
 import onnxruntime_genai as og
 
 print("Loading model...")
-model=og.Model("model", og.DeviceType.CPU)
+
+# The first argument is the name of the folder containing the model files
+model=og.Model("example-models/llama2-7b-chat-int4-cpu", og.DeviceType.CPU)
 print("Model loaded")
 tokenizer=model.create_tokenizer()
 print("Tokenizer created")

examples/llama/llama.py

@@ -2,7 +2,9 @@
 import time
 
 print("Loading model...")
-model=og.Model("model", og.DeviceType.CPU)
+
+# The first argument is the name of the folder containing the model files
+model=og.Model("example-models/llama2-7b-chat-int4-cpu", og.DeviceType.CPU)
 print("Model loaded")
 tokenizer=model.create_tokenizer()
 print("Tokenizer created")