Debarun12's picture
Create README.md
2822631 verified
|
Raw
History Blame Contribute Delete
800 Bytes

ENG_llmV03 — 95M Parameter Language Model

Built entirely from scratch in PyTorch. No pretrained weights. No from_pretrained().

Performance

Metric Value
Base PPL (WikiText-103) 24.40
GPT-3 Small PPL (reference) 26.0
Fine-tuned PPL (two-stage LoRA) 20.83
Trainable params via LoRA ~1.6M (1.8%)
Training hardware RTX 5050 (8.5GB VRAM)

Architecture

  • RoPE positional encoding
  • SwiGLU activation
  • 12-layer Transformer
  • 95M parameters
  • Trained on WikiText-103 (103M tokens)

Fine-Tuning

Two-stage LoRA: R128 → merged → R64
Dataset: 355k clean QA pairs (SciQ + ELI5 + FreebaseQA)

Full Documentation

Technical documentation →

Built by Debarun Das