README.md

metadata

license: apache-2.0
language:
  - en
  - de
  - fr
  - es
  - ka
  - tw
  - ia
  - ak
metrics:
  - bleu
pipeline_tag: text-generation
library_name: adapter-transformers
tags:
  - text-generation-inference
  - chemistry
  - finance
  - legal
  - biology
  - music
  - climate
  - medical
  - code
  - art
  - merge

SpeedAI Base 125M (v1)

SpeedAI is a foundational language model developed by SpeedTech and AI Group of Companies. This is our first open-source base model — trained from scratch — representing our mission to build Africa-powered, global-scale AI systems.

🧠 Model Details

• Architecture: GPT-style Transformer (decoder-only)
• Parameters: 125M
• Layers: 12
• Hidden Size: 768
• Heads: 12
• Context Length: 2048
• Vocabulary Size: 50,257 (custom tokenizer)

📦 Files

• config.json: Model architecture
• tokenizer/: Trained tokenizer folder
• pytorch_model.bin: Model weights (to be uploaded)
• generation_config.json: Generation settings

🌍 About Speed AI

Speed AI is part of SpeedTech Group, a technology powerhouse including:

• 💸 SpeedPay (Fintech & payments)
• 🎧 SpeedStream (Streaming platform)
• 🎵 SpeedMusic (Music distribution)
• ⚡ SpeedUp (Productivity & automation tools)

Learn more: https://speedtech.group

🧠 License

MIT — Open, transparent, and free to use (v1).

---

We’re building from scratch, not fine-tuning someone else’s brain. Join us as we shape the future of real-time, human-level AI from the ground up.