Daemontatox commited on
Commit
f637502
·
verified ·
1 Parent(s): 547b264

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +5 -5
app.py CHANGED
@@ -20,7 +20,7 @@ from transformers import (
20
  )
21
 
22
  # Configuration Constants
23
- MODEL_ID = "Daemontatox/Phi-4-COT"
24
 
25
 
26
  # Understand]: Analyze the question to identify key details and clarify the goal.
@@ -139,9 +139,9 @@ h3 {
139
  def initialize_model():
140
  """Initialize the model with appropriate configurations"""
141
  quantization_config = BitsAndBytesConfig(
142
- load_in_8bit=True,
143
- bnb_8bit_compute_dtype=torch.bfloat16,
144
- bnb_8bit_use_double_quant=True
145
  )
146
 
147
  tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
@@ -153,7 +153,7 @@ def initialize_model():
153
  torch_dtype=torch.float16,
154
  device_map="cuda",
155
  attn_implementation="flash_attention_2",
156
- #quantization_config=quantization_config
157
 
158
  )
159
 
 
20
  )
21
 
22
  # Configuration Constants
23
+ MODEL_ID = "Daemontatox/PathFinderAi3.0"
24
 
25
 
26
  # Understand]: Analyze the question to identify key details and clarify the goal.
 
139
  def initialize_model():
140
  """Initialize the model with appropriate configurations"""
141
  quantization_config = BitsAndBytesConfig(
142
+ load_in_4bit=True,
143
+ bnb_4bit_compute_dtype=torch.bfloat16,
144
+ bnb_4bit_use_double_quant=True
145
  )
146
 
147
  tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
 
153
  torch_dtype=torch.float16,
154
  device_map="cuda",
155
  attn_implementation="flash_attention_2",
156
+ quantization_config=quantization_config
157
 
158
  )
159