Di Zhang
commited on
Update app.py
Browse files
app.py
CHANGED
@@ -24,7 +24,6 @@ model = AutoModelForCausalLM.from_pretrained(
|
|
24 |
DESCRIPTION = '''
|
25 |
# SimpleBerry/LLaMA-O1-Supervised-1129 | Optimized for Streaming and Hugging Face Zero Space.
|
26 |
This model is experimental and focused on advancing AI reasoning capabilities.
|
27 |
-
|
28 |
**To start a new chat**, click "clear" and begin a fresh dialogue.
|
29 |
'''
|
30 |
|
@@ -38,12 +37,16 @@ def llama_o1_template(data):
|
|
38 |
text = template.format(content=data)
|
39 |
return text
|
40 |
|
41 |
-
|
42 |
@spaces.GPU
|
43 |
-
def
|
44 |
-
|
|
|
|
|
|
|
|
|
|
|
45 |
**inputs,
|
46 |
-
|
47 |
temperature=temperature,
|
48 |
top_p=top_p,
|
49 |
do_sample=True,
|
@@ -51,19 +54,10 @@ def gen_one_token(inputs,temperature,top_p):
|
|
51 |
pad_token_id=tokenizer.eos_token_id,
|
52 |
return_dict_in_generate=True,
|
53 |
output_scores=False
|
54 |
-
)
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
def generate_text(message, history, max_tokens=512, temperature=0.9, top_p=0.95):
|
59 |
-
input_text = llama_o1_template(message)
|
60 |
-
for i in range(max_tokens):
|
61 |
-
inputs = tokenizer(input_text, return_tensors="pt").to(accelerator.device)
|
62 |
-
output = gen_one_token(inputs,temperature,top_p)
|
63 |
-
# Return text with special tokens included
|
64 |
-
generated_text = tokenizer.decode(output, skip_special_tokens=False)
|
65 |
-
input_text += generated_text
|
66 |
-
yield generated_text
|
67 |
|
68 |
with gr.Blocks() as demo:
|
69 |
gr.Markdown(DESCRIPTION)
|
@@ -89,4 +83,4 @@ with gr.Blocks() as demo:
|
|
89 |
gr.Markdown(LICENSE)
|
90 |
|
91 |
if __name__ == "__main__":
|
92 |
-
demo.launch()
|
|
|
24 |
DESCRIPTION = '''
|
25 |
# SimpleBerry/LLaMA-O1-Supervised-1129 | Optimized for Streaming and Hugging Face Zero Space.
|
26 |
This model is experimental and focused on advancing AI reasoning capabilities.
|
|
|
27 |
**To start a new chat**, click "clear" and begin a fresh dialogue.
|
28 |
'''
|
29 |
|
|
|
37 |
text = template.format(content=data)
|
38 |
return text
|
39 |
|
|
|
40 |
@spaces.GPU
|
41 |
+
def generate_text(message, history, max_tokens=512, temperature=0.9, top_p=0.95):
|
42 |
+
input_text = llama_o1_template(message)
|
43 |
+
inputs = tokenizer(input_text, return_tensors="pt").to(accelerator.device)
|
44 |
+
|
45 |
+
# Stream generation, token by token
|
46 |
+
with torch.no_grad():
|
47 |
+
for output in model.generate(
|
48 |
**inputs,
|
49 |
+
max_length=max_tokens,
|
50 |
temperature=temperature,
|
51 |
top_p=top_p,
|
52 |
do_sample=True,
|
|
|
54 |
pad_token_id=tokenizer.eos_token_id,
|
55 |
return_dict_in_generate=True,
|
56 |
output_scores=False
|
57 |
+
):
|
58 |
+
# Return text with special tokens included
|
59 |
+
generated_text = tokenizer.decode(output, skip_special_tokens=False)
|
60 |
+
yield generated_text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
61 |
|
62 |
with gr.Blocks() as demo:
|
63 |
gr.Markdown(DESCRIPTION)
|
|
|
83 |
gr.Markdown(LICENSE)
|
84 |
|
85 |
if __name__ == "__main__":
|
86 |
+
demo.launch()
|