Fix model name from InternVL2-8B to InternVL2
Browse files
app.py
CHANGED
@@ -104,7 +104,7 @@ def create_in_memory_flash_attn_mock():
|
|
104 |
# Add attributes used by transformers checks
|
105 |
flash_attn.__version__ = "0.0.0-mocked"
|
106 |
|
107 |
-
# Create common submodules -
|
108 |
for submodule in ['flash_attn.flash_attn_interface', 'flash_attn.flash_attn_triton']:
|
109 |
parts = submodule.split('.')
|
110 |
parent_name = '.'.join(parts[:-1])
|
@@ -170,7 +170,8 @@ def load_model():
|
|
170 |
print("\nLoading InternVL2 model...")
|
171 |
|
172 |
# Load the model and tokenizer
|
173 |
-
|
|
|
174 |
|
175 |
# Print downloading status
|
176 |
print("Downloading model shards. This may take some time...")
|
|
|
104 |
# Add attributes used by transformers checks
|
105 |
flash_attn.__version__ = "0.0.0-mocked"
|
106 |
|
107 |
+
# Create common submodules - without 'parent' parameter
|
108 |
for submodule in ['flash_attn.flash_attn_interface', 'flash_attn.flash_attn_triton']:
|
109 |
parts = submodule.split('.')
|
110 |
parent_name = '.'.join(parts[:-1])
|
|
|
170 |
print("\nLoading InternVL2 model...")
|
171 |
|
172 |
# Load the model and tokenizer
|
173 |
+
# FIXED: Corrected model name from InternVL2-8B to InternVL2
|
174 |
+
model_path = "OpenGVLab/InternVL2"
|
175 |
|
176 |
# Print downloading status
|
177 |
print("Downloading model shards. This may take some time...")
|