Update app.py
Browse files
app.py
CHANGED
@@ -55,16 +55,17 @@ repo_name = "IlyaGusev/saiga_13b_lora_llamacpp"
|
|
55 |
model_name = "ggml-model-q4_1.bin"
|
56 |
embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
|
57 |
|
58 |
-
snapshot_download(repo_id=repo_name, local_dir="
|
|
|
59 |
|
60 |
model = Llama(
|
61 |
-
model_path=model_name,
|
62 |
n_ctx=2000,
|
63 |
n_parts=1,
|
64 |
)
|
65 |
|
66 |
max_new_tokens = 1500
|
67 |
-
embeddings = HuggingFaceEmbeddings(model_name=
|
68 |
|
69 |
def get_uuid():
|
70 |
return str(uuid4())
|
|
|
55 |
model_name = "ggml-model-q4_1.bin"
|
56 |
embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
|
57 |
|
58 |
+
snapshot_download(repo_id=repo_name, local_dir="/data/saiga_13b_lora_llamacpp", allow_patterns=model_name)
|
59 |
+
snapshot_download(repo_id=embedder_name, local_dir="/data/paraphrase-multilingual-mpnet-base-v2")
|
60 |
|
61 |
model = Llama(
|
62 |
+
model_path=f"/data/saiga_13b_lora_llamacpp/{model_name}",
|
63 |
n_ctx=2000,
|
64 |
n_parts=1,
|
65 |
)
|
66 |
|
67 |
max_new_tokens = 1500
|
68 |
+
embeddings = HuggingFaceEmbeddings(model_name="/data/paraphrase-multilingual-mpnet-base-v2")
|
69 |
|
70 |
def get_uuid():
|
71 |
return str(uuid4())
|