GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Ronin48LLC
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Ronin48LLC
Model Tree
meta-llama/Llama-3.3-70B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Field | Value |
|---|---|
| Base Model | meta-llama/Llama-3.3-70B-Instruct |
| Adapter Type | LoRA (QLoRA) |
| LoRA Rank / Alpha / Dropout | 64 / 128 / 0.05 |
| Target Modules | q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj |
| Task | Fireground size-up, incident command, hazmat and rescue decision support |
| Training Method | QLoRA supervised fine-tuning via SFTTrainer (trl) |
| License | MIT (adapter weights) |
| Maintainer | Ronin48LLC |
For trained fire service personnel:
BRUNO is a tactical decision support tool, not a replacement for your training, your officer, or your IC. When in doubt, call your IC.
Read LIMITATIONS.md before deploying in any operational context. In short:
This is a PEFT LoRA adapter. Load it on top of the base model:
python
from peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer base_model = AutoModelForCausalLM.from_pretrained( "meta-llama/Llama-3.3-70B-Instruct", load_in_4bit=True, device_map="auto",)model = PeftModel.from_pretrained(base_model, "Ronin48LLC/bruno-lora-adapter")tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.3-70B-Instruct")Note: Access to the base model requires accepting Meta's license on Hugging Face.
The base model with the BRUNO system prompt and no fine-tuning is a workable baseline; see the source repository's Quick Start.
adapter_config.json)training_args.bin is preserved in this repository; the step count and loss curve were not recorded in the upload| Source | Description | License |
|---|---|---|
| NFPA 1 / NFPA 101 / NFPA 13 | Fire code, life safety, suppression systems | Public reference |
| IFSTA Essentials / Fire Officer | Core firefighter and officer curriculum | Public guidelines |
| NIOSH Fire Fighter Fatality Reports | Line-of-duty death investigations | Public Domain |
| NIST Fire Research | Structural fire behavior research, UL studies | Public Domain |
| DOT Emergency Response Guidebook | Hazardous materials emergency response | Public Domain |
| ICS-100 through ICS-400 | FEMA NIMS/ICS training materials | Public Domain |
| FEMA/USFA Fire Data | National fire statistics, incident data | Public Domain |
| Synthetic Scenarios | AI-generated fireground and incident scenarios | Proprietary |
BRUNO, BONES, and SELMA share scenes constantly — consult the appropriate model for each domain.
| Model | Domain | Use when… |
|---|---|---|
| BRUNO (this adapter) | Fire service — company officer / IC | Fireground tactics, size-up, hazmat, extrication, water supply, ICS |
| BONES | EMS — EMR / EMT / AEMT / Paramedic | Patient assessment, treatment protocols, drug dosing, triage, transport |
| SELMA | Law enforcement | Criminal statute identification, charge elements, constitutional flags |
| Shared scene | Primary | Support |
|---|---|---|
| Structure fire with casualties | BRUNO (fireground ops) | BONES (patient care) |
| Vehicle accident with entrapment | BRUNO (extrication) | BONES (care during extrication) |
| Hazmat with patient exposures | BRUNO (mitigation, decon zone) | BONES (patient decon and treatment) |
| Mass casualty incident | BONES (triage, treatment) | BRUNO (ICS, sectors) + SELMA (criminal nexus) |
| Arson investigation | BRUNO (origin/cause, fire behavior) | SELMA (arson statutes) + BONES (patient care) |
| Wildland/WUI with structure threat | BRUNO (tactical operations) | BONES (civilian casualties) |
| Confined-space rescue | BRUNO (rescue ops, atmospheric monitoring) | BONES (patient extraction and care) |
ABBY (forensics) operates independently of the first-responder suite; SELMA pairs with ATTICUS on the legal side.
github.com/CryptoJones/BRUNO — data collection, QLoRA pipeline, synthetic-scenario generator.
Adapter weights: MIT — Copyright 2026 Ronin 48, LLC. Base model weights are subject to the Meta Llama 3.3 Community License.
from peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizer base_model = AutoModelForCausalLM.from_pretrained( "meta-llama/Llama-3.3-70B-Instruct", load_in_4bit=True, device_map="auto",)model = PeftModel.from_pretrained(base_model, "Ronin48LLC/bruno-lora-adapter")tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-3.3-70B-Instruct")