Update app.py
Browse files
app.py
CHANGED
@@ -27,8 +27,13 @@ app.logger.setLevel(logging.DEBUG) # Set the desired logging level
|
|
27 |
#model_path = "../models/model-q4_K.gguf" # Replace with the actual model path
|
28 |
#model_name = "model/ggml-model-q4_K.gguf"
|
29 |
|
30 |
-
repo_name = "IlyaGusev/saiga2_13b_gguf"
|
31 |
-
model_name = "model-q4_K.gguf"
|
|
|
|
|
|
|
|
|
|
|
32 |
|
33 |
snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
|
34 |
|
@@ -39,7 +44,8 @@ model = Llama(
|
|
39 |
#n_batch=100,
|
40 |
logits_all=True,
|
41 |
#n_threads=12,
|
42 |
-
verbose=True
|
|
|
43 |
)
|
44 |
|
45 |
|
|
|
27 |
#model_path = "../models/model-q4_K.gguf" # Replace with the actual model path
|
28 |
#model_name = "model/ggml-model-q4_K.gguf"
|
29 |
|
30 |
+
#repo_name = "IlyaGusev/saiga2_13b_gguf"
|
31 |
+
#model_name = "model-q4_K.gguf"
|
32 |
+
|
33 |
+
repo_name = "IlyaGusev/saiga2_70b_gguf"
|
34 |
+
model_name = "ggml-model-q4_1.gguf"
|
35 |
+
|
36 |
+
|
37 |
|
38 |
snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
|
39 |
|
|
|
44 |
#n_batch=100,
|
45 |
logits_all=True,
|
46 |
#n_threads=12,
|
47 |
+
verbose=True,
|
48 |
+
n_gqa=8 #must be set for 70b models
|
49 |
)
|
50 |
|
51 |
|