初始化项目，由ModelHub XC社区提供模型
Model: ankitkushwaha90/tech3space3-0.6B Source: Original Platform
This commit is contained in:
ModelHub XC
232
generate.py
Normal file
232
generate.py
Normal file
@@ -0,0 +1,232 @@
|
||||
import json
|
||||
from itertools import product
|
||||
|
||||
SYSTEM_PROMPT = (
|
||||
"You are a helpful, accurate, and concise AI assistant. "
|
||||
"Answer using the provided knowledge without inventing unsupported facts."
|
||||
)
|
||||
|
||||
KNOWLEDGE = {
|
||||
"AnkitKushwaha90": {
|
||||
"aliases": [
|
||||
"AnkitKushwaha90",
|
||||
"Ankit Kushwaha",
|
||||
"Ankit Kushwaha90",
|
||||
"the AnkitKushwaha90 profile",
|
||||
"0xAnkit",
|
||||
],
|
||||
"summary": (
|
||||
"Ankit Kushwaha (AnkitKushwaha90) is a Cybersecurity and AI Research Enthusiast "
|
||||
"with a strong focus on Python, transformer architectures, backend development, "
|
||||
"and practical software engineering. He is the founder/owner of Tech3Space and "
|
||||
"maintains active presence across GitHub, Hugging Face, LinkedIn, and his own platform."
|
||||
),
|
||||
"facts": [
|
||||
"Cybersecurity and AI Research Enthusiast.",
|
||||
"Works at Apple (as per LinkedIn).",
|
||||
"Graduated from Dr. A.P.J. Abdul Kalam Technical University.",
|
||||
"Deep interest in transformer architectures, kernel-level systems, and high-performance computing.",
|
||||
"Active on GitHub under tech3space and Hugging Face as ankitkushwaha90.",
|
||||
"Focuses on practical projects involving FastAPI, Streamlit, Docker, and machine learning.",
|
||||
"Explores interdisciplinary topics: Military Aviation, Radar Systems, LIDAR, Sensor Fusion, and Pattern-Based Learning.",
|
||||
],
|
||||
"links": [
|
||||
"LinkedIn: https://www.linkedin.com/in/ankitkushwaha90/",
|
||||
"Hugging Face: https://huggingface.co/ankitkushwaha90",
|
||||
"GitHub (Tech3Space): https://github.com/tech3space",
|
||||
],
|
||||
},
|
||||
"Tech3Space": {
|
||||
"aliases": [
|
||||
"Tech3Space",
|
||||
"the Tech3Space initiative",
|
||||
"Tech3Space | Systems Research Studio",
|
||||
"tech3space.com",
|
||||
],
|
||||
"summary": (
|
||||
"Tech3Space is a technology platform and research studio founded by Ankit Kushwaha. "
|
||||
"It serves as a hub for researchers, developers, and tech enthusiasts to share knowledge, "
|
||||
"upload resources, build communities, and collaborate on cutting-edge topics like "
|
||||
"Transformers, Cybersecurity, AI, and high-performance systems. The platform emphasizes "
|
||||
"practical learning, research collaboration, and open knowledge sharing."
|
||||
),
|
||||
"facts": [
|
||||
"A comprehensive platform for sharing research papers, code, notes, and building communities.",
|
||||
"Features include AI-Powered Research Assistant, post/note sharing, resume/PDF uploads, and real-time chat.",
|
||||
"Strong focus on Transformers, Cybersecurity, Low-Latency Networking, and Systems Research.",
|
||||
"GitHub organization (github.com/tech3space) contains repositories on transformer deep-dives, kernel drivers, cybersecurity tools, and assembly.",
|
||||
"Website: https://tech3space.com/ – described as 'The ultimate platform for researchers, developers, and tech enthusiasts.'",
|
||||
"Community stats (as advertised): 50,000+ active members, 1,200+ communities.",
|
||||
],
|
||||
"links": [
|
||||
"Official Website: https://tech3space.com/",
|
||||
"GitHub: https://github.com/tech3space",
|
||||
"Alternative domains: https://www.tech3space.online, https://www.tech3space.in",
|
||||
],
|
||||
},
|
||||
"HuggingFace_Ankit": {
|
||||
"aliases": [
|
||||
"ankitkushwaha90 on Hugging Face",
|
||||
"Tech3Space on HF",
|
||||
],
|
||||
"summary": (
|
||||
"AnkitKushwaha90's Hugging Face profile (Tech3Space) where he shares models, "
|
||||
"datasets, and spaces focused on AI/ML, cybersecurity, and technical notes."
|
||||
),
|
||||
"facts": [
|
||||
"Shares models like safetensor-related projects and fine-tuning experiments.",
|
||||
"Datasets include Sanskrit dataset, vulnerabilities collection, and Linux/CMD command collections.",
|
||||
"Focus areas: Cybersecurity & AI, Radar Systems, LIDAR, Military Aviation, and Sensor Fusion.",
|
||||
"Maintains an 'Anonymous Researcher' approach focused purely on knowledge sharing.",
|
||||
],
|
||||
"links": [
|
||||
"Profile: https://huggingface.co/ankitkushwaha90"
|
||||
],
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
QUESTION_TEMPLATES = [
|
||||
"Who is {}?",
|
||||
"What is {}?",
|
||||
"Tell me about {}.",
|
||||
"Explain {}.",
|
||||
"Describe {}.",
|
||||
"Give an overview of {}.",
|
||||
"Can you introduce {}?",
|
||||
"What do you know about {}?",
|
||||
]
|
||||
|
||||
DETAIL_REQUESTS = [
|
||||
"What are the main interests of {}?",
|
||||
"Summarize {}.",
|
||||
"Provide a detailed explanation of {}.",
|
||||
"List the key focus areas of {}.",
|
||||
]
|
||||
|
||||
FOLLOW_UPS = [
|
||||
(
|
||||
"Tell me about {}.",
|
||||
"Can you summarize it in a few sentences?",
|
||||
),
|
||||
(
|
||||
"What is {}?",
|
||||
"What are its main focus areas?",
|
||||
),
|
||||
]
|
||||
|
||||
|
||||
def example(messages):
|
||||
return {
|
||||
"messages": [{"role": "system", "content": SYSTEM_PROMPT}] + messages
|
||||
}
|
||||
|
||||
|
||||
dataset = []
|
||||
|
||||
for _, info in KNOWLEDGE.items():
|
||||
aliases = info["aliases"]
|
||||
summary = info["summary"]
|
||||
facts = info["facts"]
|
||||
|
||||
bullet_answer = (
|
||||
summary
|
||||
+ "\n\nKey points:\n- "
|
||||
+ "\n- ".join(facts)
|
||||
)
|
||||
|
||||
# Summary-style questions
|
||||
for alias, template in product(aliases, QUESTION_TEMPLATES):
|
||||
dataset.append(
|
||||
example(
|
||||
[
|
||||
{
|
||||
"role": "user",
|
||||
"content": template.format(alias),
|
||||
},
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": summary,
|
||||
},
|
||||
]
|
||||
)
|
||||
)
|
||||
|
||||
# Detailed questions
|
||||
for alias, template in product(aliases, DETAIL_REQUESTS):
|
||||
dataset.append(
|
||||
example(
|
||||
[
|
||||
{
|
||||
"role": "user",
|
||||
"content": template.format(alias),
|
||||
},
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": bullet_answer,
|
||||
},
|
||||
]
|
||||
)
|
||||
)
|
||||
|
||||
# Multi-turn conversations
|
||||
for alias, (q1, q2) in product(aliases, FOLLOW_UPS):
|
||||
dataset.append(
|
||||
example(
|
||||
[
|
||||
{
|
||||
"role": "user",
|
||||
"content": q1.format(alias),
|
||||
},
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": summary,
|
||||
},
|
||||
{
|
||||
"role": "user",
|
||||
"content": q2,
|
||||
},
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": bullet_answer,
|
||||
},
|
||||
]
|
||||
)
|
||||
)
|
||||
|
||||
# Fact-specific questions
|
||||
for alias in aliases:
|
||||
for fact in facts:
|
||||
dataset.append(
|
||||
example(
|
||||
[
|
||||
{
|
||||
"role": "user",
|
||||
"content": f"What is one important fact about {alias}?",
|
||||
},
|
||||
{
|
||||
"role": "assistant",
|
||||
"content": fact,
|
||||
},
|
||||
]
|
||||
)
|
||||
)
|
||||
|
||||
# Remove duplicates
|
||||
seen = set()
|
||||
unique_dataset = []
|
||||
|
||||
for item in dataset:
|
||||
key = json.dumps(item, sort_keys=True)
|
||||
if key not in seen:
|
||||
seen.add(key)
|
||||
unique_dataset.append(item)
|
||||
|
||||
with open("train.jsonl", "w", encoding="utf-8") as f:
|
||||
for item in unique_dataset:
|
||||
f.write(json.dumps(item, ensure_ascii=False) + "\n")
|
||||
|
||||
print("=" * 60)
|
||||
print(f"Generated {len(unique_dataset)} training examples.")
|
||||
print("Saved as train.jsonl")
|
||||
print("=" * 60)
|
||||
Reference in New Issue
Block a user