CreitinGameplays
commited on
Commit
•
bc4308f
1
Parent(s):
a487b7f
Update app.py
Browse files
app.py
CHANGED
@@ -12,7 +12,7 @@ model = AutoModelForCausalLM.from_pretrained(model_name)
|
|
12 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
13 |
model.to(device)
|
14 |
|
15 |
-
@spaces.GPU(duration=
|
16 |
def generate_text(user_prompt):
|
17 |
"""Generates text using the BLOOM model from Hugging Face Transformers and removes the user prompt."""
|
18 |
# Construct the full prompt with system introduction, user prompt, and assistant role
|
@@ -24,7 +24,7 @@ def generate_text(user_prompt):
|
|
24 |
# Generate text with the complete prompt and limit the maximum length to 256 tokens
|
25 |
output = model.generate(
|
26 |
input_ids=prompt_encoded,
|
27 |
-
max_length=
|
28 |
num_beams=1,
|
29 |
num_return_sequences=1,
|
30 |
do_sample=True,
|
|
|
12 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
13 |
model.to(device)
|
14 |
|
15 |
+
@spaces.GPU(duration=90)
|
16 |
def generate_text(user_prompt):
|
17 |
"""Generates text using the BLOOM model from Hugging Face Transformers and removes the user prompt."""
|
18 |
# Construct the full prompt with system introduction, user prompt, and assistant role
|
|
|
24 |
# Generate text with the complete prompt and limit the maximum length to 256 tokens
|
25 |
output = model.generate(
|
26 |
input_ids=prompt_encoded,
|
27 |
+
max_length=1900,
|
28 |
num_beams=1,
|
29 |
num_return_sequences=1,
|
30 |
do_sample=True,
|