Spaces:
Running
on
Zero
Running
on
Zero
adds mcp server, ssrmode false , refactor the global variables
Browse files
app.py
CHANGED
@@ -5,13 +5,6 @@ from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
|
|
5 |
import torch
|
6 |
import json
|
7 |
|
8 |
-
LEAN4_DEFAULT_HEADER = (
|
9 |
-
"import Mathlib\n"
|
10 |
-
"import Aesop\n\n"
|
11 |
-
"set_option maxHeartbeats 0\n\n"
|
12 |
-
"open BigOperators Real Nat Topology Rat\n"
|
13 |
-
)
|
14 |
-
|
15 |
title = "# 🙋🏻♂️Welcome to 🌟Tonic's 🌕💉👨🏻🔬Moonshot Math"
|
16 |
|
17 |
description = """
|
@@ -39,27 +32,16 @@ joinus = """
|
|
39 |
🌟TeamTonic🌟 is always making cool demos! Join our active builder's 🛠️community 👻 [](https://discord.gg/qdfnvSPcqP) On 🤗Huggingface:[MultiTransformer](https://huggingface.co/MultiTransformer) On 🌐Github: [Tonic-AI](https://github.com/tonic-ai) & contribute to🌟 [MultiTonic](https://github.com/MultiTonic)🤗Big thanks to Yuvi Sharma and all the folks at huggingface for the community grant 🤗
|
40 |
"""
|
41 |
|
42 |
-
# Build the initial system prompt
|
43 |
SYSTEM_PROMPT = "You are an expert in mathematics and Lean 4."
|
44 |
|
45 |
-
# Helper to build a Lean4 code block
|
46 |
-
def build_formal_block(formal_statement, informal_prefix=""):
|
47 |
-
return (
|
48 |
-
f"{LEAN4_DEFAULT_HEADER}\n"
|
49 |
-
f"{informal_prefix}\n"
|
50 |
-
f"{formal_statement}"
|
51 |
-
)
|
52 |
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
return '\n'.join(lines)
|
60 |
-
return text.strip()
|
61 |
|
62 |
-
# Example problems
|
63 |
unimath1 = """Goal:
|
64 |
X : UU
|
65 |
Y : UU
|
@@ -88,6 +70,21 @@ unimath4 = """Goal:
|
|
88 |
|
89 |
additional_info_prompt = "/-Explain using mathematics-/\n"
|
90 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
examples = [
|
92 |
[unimath1, additional_info_prompt, 1234],
|
93 |
[unimath2, additional_info_prompt, 1234],
|
@@ -114,9 +111,7 @@ model_name = "AI-MO/Kimina-Prover-Distill-8B"
|
|
114 |
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
115 |
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True)
|
116 |
|
117 |
-
# Set generation config
|
118 |
model.generation_config = GenerationConfig.from_pretrained(model_name)
|
119 |
-
# Ensure pad_token_id is an integer, not a list
|
120 |
if isinstance(model.generation_config.eos_token_id, list):
|
121 |
model.generation_config.pad_token_id = model.generation_config.eos_token_id[0]
|
122 |
else:
|
@@ -125,7 +120,6 @@ model.generation_config.do_sample = True
|
|
125 |
model.generation_config.temperature = 0.6
|
126 |
model.generation_config.top_p = 0.95
|
127 |
|
128 |
-
# Initialize chat history with system prompt
|
129 |
def init_chat(formal_statement, informal_prefix):
|
130 |
user_prompt = (
|
131 |
"Think about and solve the following problem step by step in Lean 4.\n"
|
@@ -137,15 +131,12 @@ def init_chat(formal_statement, informal_prefix):
|
|
137 |
{"role": "user", "content": user_prompt}
|
138 |
]
|
139 |
|
140 |
-
# Gradio chat handler
|
141 |
@spaces.GPU
|
142 |
def chat_handler(user_message, informal_prefix, max_tokens, chat_history):
|
143 |
-
# If chat_history is empty, initialize with system and first user message
|
144 |
if not chat_history or len(chat_history) < 2:
|
145 |
chat_history = init_chat(user_message, informal_prefix)
|
146 |
display_history = [("user", user_message)]
|
147 |
else:
|
148 |
-
# Append new user message
|
149 |
chat_history.append({"role": "user", "content": user_message})
|
150 |
display_history = []
|
151 |
for msg in chat_history:
|
@@ -153,7 +144,6 @@ def chat_handler(user_message, informal_prefix, max_tokens, chat_history):
|
|
153 |
display_history.append(("user", msg["content"]))
|
154 |
elif msg["role"] == "assistant":
|
155 |
display_history.append(("assistant", msg["content"]))
|
156 |
-
# Format prompt using chat template
|
157 |
prompt = tokenizer.apply_chat_template(chat_history, tokenize=False, add_generation_prompt=True)
|
158 |
input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(model.device)
|
159 |
attention_mask = torch.ones_like(input_ids)
|
@@ -166,14 +156,10 @@ def chat_handler(user_message, informal_prefix, max_tokens, chat_history):
|
|
166 |
top_p=model.generation_config.top_p,
|
167 |
)
|
168 |
result = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
169 |
-
# Extract only the new assistant message (after the prompt)
|
170 |
new_response = result[len(prompt):].strip()
|
171 |
-
# Add assistant message to chat history
|
172 |
chat_history.append({"role": "assistant", "content": new_response})
|
173 |
display_history.append(("assistant", new_response))
|
174 |
-
# Extract Lean4 code
|
175 |
code = extract_lean4_code(new_response)
|
176 |
-
# Prepare output
|
177 |
output_data = {
|
178 |
"model_input": prompt,
|
179 |
"model_output": result,
|
@@ -184,7 +170,6 @@ def chat_handler(user_message, informal_prefix, max_tokens, chat_history):
|
|
184 |
|
185 |
def main():
|
186 |
with gr.Blocks() as demo:
|
187 |
-
# Title and Model Description
|
188 |
gr.Markdown("""# 🙋🏻♂️Welcome to 🌟Tonic's 🌕💉👨🏻🔬Moonshot Math""")
|
189 |
with gr.Row():
|
190 |
with gr.Column():
|
@@ -213,4 +198,4 @@ def main():
|
|
213 |
demo.launch()
|
214 |
|
215 |
if __name__ == "__main__":
|
216 |
-
main()
|
|
|
5 |
import torch
|
6 |
import json
|
7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
title = "# 🙋🏻♂️Welcome to 🌟Tonic's 🌕💉👨🏻🔬Moonshot Math"
|
9 |
|
10 |
description = """
|
|
|
32 |
🌟TeamTonic🌟 is always making cool demos! Join our active builder's 🛠️community 👻 [](https://discord.gg/qdfnvSPcqP) On 🤗Huggingface:[MultiTransformer](https://huggingface.co/MultiTransformer) On 🌐Github: [Tonic-AI](https://github.com/tonic-ai) & contribute to🌟 [MultiTonic](https://github.com/MultiTonic)🤗Big thanks to Yuvi Sharma and all the folks at huggingface for the community grant 🤗
|
33 |
"""
|
34 |
|
|
|
35 |
SYSTEM_PROMPT = "You are an expert in mathematics and Lean 4."
|
36 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
+
LEAN4_DEFAULT_HEADER = (
|
39 |
+
"import Mathlib\n"
|
40 |
+
"import Aesop\n\n"
|
41 |
+
"set_option maxHeartbeats 0\n\n"
|
42 |
+
"open BigOperators Real Nat Topology Rat\n"
|
43 |
+
)
|
|
|
|
|
44 |
|
|
|
45 |
unimath1 = """Goal:
|
46 |
X : UU
|
47 |
Y : UU
|
|
|
70 |
|
71 |
additional_info_prompt = "/-Explain using mathematics-/\n"
|
72 |
|
73 |
+
def build_formal_block(formal_statement, informal_prefix=""):
|
74 |
+
return (
|
75 |
+
f"{LEAN4_DEFAULT_HEADER}\n"
|
76 |
+
f"{informal_prefix}\n"
|
77 |
+
f"{formal_statement}"
|
78 |
+
)
|
79 |
+
|
80 |
+
def extract_lean4_code(text):
|
81 |
+
code_block = re.search(r"```lean4(.*?)(```|$)", text, re.DOTALL)
|
82 |
+
if code_block:
|
83 |
+
code = code_block.group(1)
|
84 |
+
lines = [line for line in code.split('\n') if line.strip()]
|
85 |
+
return '\n'.join(lines)
|
86 |
+
return text.strip()
|
87 |
+
|
88 |
examples = [
|
89 |
[unimath1, additional_info_prompt, 1234],
|
90 |
[unimath2, additional_info_prompt, 1234],
|
|
|
111 |
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
112 |
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto", trust_remote_code=True)
|
113 |
|
|
|
114 |
model.generation_config = GenerationConfig.from_pretrained(model_name)
|
|
|
115 |
if isinstance(model.generation_config.eos_token_id, list):
|
116 |
model.generation_config.pad_token_id = model.generation_config.eos_token_id[0]
|
117 |
else:
|
|
|
120 |
model.generation_config.temperature = 0.6
|
121 |
model.generation_config.top_p = 0.95
|
122 |
|
|
|
123 |
def init_chat(formal_statement, informal_prefix):
|
124 |
user_prompt = (
|
125 |
"Think about and solve the following problem step by step in Lean 4.\n"
|
|
|
131 |
{"role": "user", "content": user_prompt}
|
132 |
]
|
133 |
|
|
|
134 |
@spaces.GPU
|
135 |
def chat_handler(user_message, informal_prefix, max_tokens, chat_history):
|
|
|
136 |
if not chat_history or len(chat_history) < 2:
|
137 |
chat_history = init_chat(user_message, informal_prefix)
|
138 |
display_history = [("user", user_message)]
|
139 |
else:
|
|
|
140 |
chat_history.append({"role": "user", "content": user_message})
|
141 |
display_history = []
|
142 |
for msg in chat_history:
|
|
|
144 |
display_history.append(("user", msg["content"]))
|
145 |
elif msg["role"] == "assistant":
|
146 |
display_history.append(("assistant", msg["content"]))
|
|
|
147 |
prompt = tokenizer.apply_chat_template(chat_history, tokenize=False, add_generation_prompt=True)
|
148 |
input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(model.device)
|
149 |
attention_mask = torch.ones_like(input_ids)
|
|
|
156 |
top_p=model.generation_config.top_p,
|
157 |
)
|
158 |
result = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
|
|
159 |
new_response = result[len(prompt):].strip()
|
|
|
160 |
chat_history.append({"role": "assistant", "content": new_response})
|
161 |
display_history.append(("assistant", new_response))
|
|
|
162 |
code = extract_lean4_code(new_response)
|
|
|
163 |
output_data = {
|
164 |
"model_input": prompt,
|
165 |
"model_output": result,
|
|
|
170 |
|
171 |
def main():
|
172 |
with gr.Blocks() as demo:
|
|
|
173 |
gr.Markdown("""# 🙋🏻♂️Welcome to 🌟Tonic's 🌕💉👨🏻🔬Moonshot Math""")
|
174 |
with gr.Row():
|
175 |
with gr.Column():
|
|
|
198 |
demo.launch()
|
199 |
|
200 |
if __name__ == "__main__":
|
201 |
+
main(ssr_mode=False, mcp_server=True)
|