Spaces:
Runtime error
Runtime error
FlawedLLM
commited on
Commit
•
9d7277a
1
Parent(s):
dcde33a
Update app.py
Browse files
app.py
CHANGED
@@ -7,11 +7,11 @@ from peft import PeftModel, PeftConfig
|
|
7 |
|
8 |
|
9 |
tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/BhashiniLLM")
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
# model = AutoModelForCausalLM.from_pretrained("FlawedLLM/BhashiniLLM",
|
16 |
# device_map="auto",
|
17 |
# quantization_config=quantization_config,
|
@@ -35,7 +35,7 @@ tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/BhashiniLLM")
|
|
35 |
|
36 |
|
37 |
config = PeftConfig.from_pretrained("FlawedLLM/BhashiniLLM")
|
38 |
-
base_model = AutoModelForCausalLM.from_pretrained("unsloth/llama-3-8b-bnb-4bit", device_map='auto')
|
39 |
model = PeftModel.from_pretrained(base_model, "FlawedLLM/BhashiniLLM")
|
40 |
|
41 |
|
|
|
7 |
|
8 |
|
9 |
tokenizer = AutoTokenizer.from_pretrained("FlawedLLM/BhashiniLLM")
|
10 |
+
quantization_config = BitsAndBytesConfig(
|
11 |
+
load_in_4bit=True,
|
12 |
+
bnb_4bit_use_double_quant=True,
|
13 |
+
bnb_4bit_quant_type="nf4",
|
14 |
+
bnb_4bit_compute_dtype=torch.float16)
|
15 |
# model = AutoModelForCausalLM.from_pretrained("FlawedLLM/BhashiniLLM",
|
16 |
# device_map="auto",
|
17 |
# quantization_config=quantization_config,
|
|
|
35 |
|
36 |
|
37 |
config = PeftConfig.from_pretrained("FlawedLLM/BhashiniLLM")
|
38 |
+
base_model = AutoModelForCausalLM.from_pretrained("unsloth/llama-3-8b-bnb-4bit", device_map='auto', quantization_config=quantization_config)
|
39 |
model = PeftModel.from_pretrained(base_model, "FlawedLLM/BhashiniLLM")
|
40 |
|
41 |
|