fix #2392 (#2393)

As noted in
https://github.com/ggerganov/llama.cpp/blob/master/migrate-ggml-2023-03-30-pr613.py,

Authors from `llama.cpp` caused a breaking change to the file format on
2023-03-30 in: ggml-org/llama.cpp#613

Therefore, we need further use `migrate-ggml-2023-03-30-pr613.py` to
convert the llama model.

Author: ShoufaChen

tests/integration_tests/embeddings/test_llamacpp.py

@@ -14,17 +14,22 @@ def get_model() -> str:
     model_url = "https://huggingface.co/Sosaka/Alpaca-native-4bit-ggml/resolve/main/ggml-alpaca-7b-q4.bin"
     tokenizer_url = "https://huggingface.co/decapoda-research/llama-7b-hf/resolve/main/tokenizer.model"
     conversion_script = "https://github.com/ggerganov/llama.cpp/raw/master/convert-unversioned-ggml-to-ggml.py"
+    migrate_script = "https://github.com/ggerganov/llama.cpp/raw/master/migrate-ggml-2023-03-30-pr613.py"
     local_filename = model_url.split("/")[-1]
+    local_filename_ggjt = local_filename.split('.')[0] + '-ggjt.' + local_filename.split('.')[1]
 
     if not os.path.exists("convert-unversioned-ggml-to-ggml.py"):
         urlretrieve(conversion_script, "convert-unversioned-ggml-to-ggml.py")
+    if not os.path.exists("migrate-ggml-2023-03-30-pr613.py"):
+        urlretrieve(migrate_script, "migrate-ggml-2023-03-30-pr613.py")
     if not os.path.exists("tokenizer.model"):
         urlretrieve(tokenizer_url, "tokenizer.model")
     if not os.path.exists(local_filename):
         urlretrieve(model_url, local_filename)
-        os.system("python convert-unversioned-ggml-to-ggml.py . tokenizer.model")
+        os.system(f"python convert-unversioned-ggml-to-ggml.py . tokenizer.model")
+        os.system(f"python migrate-ggml-2023-03-30-pr613.py {local_filename} {local_filename_ggjt}")
 
-    return local_filename
+    return local_filename_ggjt
 
 
 def test_llamacpp_embedding_documents() -> None:

tests/integration_tests/llms/test_llamacpp.py

@@ -13,17 +13,22 @@ def get_model() -> str:
     model_url = "https://huggingface.co/Sosaka/Alpaca-native-4bit-ggml/resolve/main/ggml-alpaca-7b-q4.bin"
     tokenizer_url = "https://huggingface.co/decapoda-research/llama-7b-hf/resolve/main/tokenizer.model"
     conversion_script = "https://github.com/ggerganov/llama.cpp/raw/master/convert-unversioned-ggml-to-ggml.py"
+    migrate_script = "https://github.com/ggerganov/llama.cpp/raw/master/migrate-ggml-2023-03-30-pr613.py"
     local_filename = model_url.split("/")[-1]
+    local_filename_ggjt = local_filename.split('.')[0] + '-ggjt.' + local_filename.split('.')[1]
 
     if not os.path.exists("convert-unversioned-ggml-to-ggml.py"):
         urlretrieve(conversion_script, "convert-unversioned-ggml-to-ggml.py")
+    if not os.path.exists("migrate-ggml-2023-03-30-pr613.py"):
+        urlretrieve(migrate_script, "migrate-ggml-2023-03-30-pr613.py")
     if not os.path.exists("tokenizer.model"):
         urlretrieve(tokenizer_url, "tokenizer.model")
     if not os.path.exists(local_filename):
         urlretrieve(model_url, local_filename)
         os.system(f"python convert-unversioned-ggml-to-ggml.py . tokenizer.model")
+        os.system(f"python migrate-ggml-2023-03-30-pr613.py {local_filename} {local_filename_ggjt}")
 
-    return local_filename
+    return local_filename_ggjt
 
 
 def test_llamacpp_inference() -> None: