README.md

metadata

language:
  - en
license: mit
tags:
  - devops
  - cicd
  - docker
  - kubernetes
  - infrastructure
  - slm
  - llama-style
  - rope
  - 5m-context
  - from-scratch
  - 1b-params
pipeline_tag: text-generation

DevOps Engineer-SLM: Role-Based Small Language Model

A LLaMA-style transformer (~989.8M params, ~0.99B) trained from scratch for the DevOps Engineer role. Supports up to 5M token context via RoPE with gradient checkpointing.

Architecture

Component	Value
Architecture	LLaMA-style (RoPE + RMSNorm + SwiGLU)
Parameters	989.8M (0.99B)
Layers	32
Heads	20
Embedding	1600
Max Context	5,000,000 tokens
Max Output	5,000,000 tokens
Vocab	2,107 BPE
Model Size	~4 GB (fp32)

Training

• Best eval loss: 2.5998684406280517
• Trained with gradient checkpointing on Apple M4 (MPS)
• 3 epochs, batch_size=1, grad_accum=16

Usage

from huggingface_hub import hf_hub_download
from tokenizers import Tokenizer

model_path = hf_hub_download("sathishphdai/devops-engineer-slm-5m", "model.safetensors")
tokenizer_path = hf_hub_download("sathishphdai/devops-engineer-slm-5m", "devops_engineer_tokenizer.json")
tokenizer = Tokenizer.from_file(tokenizer_path)