Irfanuruchi/OLMo-2-1124-7B-BuildEng-GGUF overview
OLMo 2 1124 7B BuildEng GGUF GGUF builds of Irfanuruchi/OLMo 2 1124 7B BuildEng . The model is a 7B OLMo 2 derivative fine tuned for building engineering. File…
Runs locally from ~4.16 GB disk (8 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| OLMo-2-1124-7B-BuildEng-BF16.gguf | GGUF | BF16 | 13.60 GB | Download |
| OLMo-2-1124-7B-BuildEng-Q4_K_M.gguf | GGUF | Q4_K_M | 4.16 GB | Download |
| OLMo-2-1124-7B-BuildEng-Q5_K_M.gguf | GGUF | Q5_K_M | 4.85 GB | Download |
| OLMo-2-1124-7B-BuildEng-Q6_K.gguf | GGUF | Q6_K | 5.58 GB | Download |
| OLMo-2-1124-7B-BuildEng-Q8_0.gguf | GGUF | Q8_0 | 7.23 GB | Download |
Model Details
| Model ID | Irfanuruchi/OLMo-2-1124-7B-BuildEng-GGUF |
|---|---|
| Author | Irfanuruchi |
| Pipeline | text-generation |
| License | apache-2.0 |
| Base model | Irfanuruchi/OLMo-2-1124-7B-BuildEng |
| Last modified | 2026-10-13T15:04:28.000Z |
Model README
---
license: apache-2.0
language:
- en
pipeline_tag: text-generation
base_model: Irfanuruchi/OLMo-2-1124-7B-BuildEng
tags:
- gguf
- llama.cpp
- olmo2
- building-engineering
- civil-engineering
- structural-engineering
---
OLMo-2-1124-7B-BuildEng-GGUF
GGUF builds of Irfanuruchi/OLMo-2-1124-7B-BuildEng.
The model is a 7B OLMo 2 derivative fine-tuned for building engineering.
Files
| Format | File size |
|---|---:|
| BF16 | 14.60 GB |
| Q8_0 | 7.76 GB |
| Q6_K | 5.99 GB |
| Q5_K_M | 5.21 GB |
| Q4_K_M | 4.47 GB |
Q5_K_M is the practical default I recommend from this release.
It was 5.21 GB and had the highest observed aggregate score in the blinded quantization-retention check. That result should not be interpreted as evidence that Q5_K_M is inherently better than BF16. Small numerical changes can alter greedy decoding trajectories, and the score differences in this suite were not statistically significant.
Q4_K_M is the smallest build. In this particular check it achieved 162 points versus 171 for BF16, or about 94.7% of the BF16 aggregate score.
Blind qualification
The same 16 Building Engineering prompts and generation settings were used for all five formats.
| Format | Aggregate score | Mean / 20 |
|---|---:|---:|
| Q5_K_M | 182 / 320 | 11.375 |
| Q8_0 | 180 / 320 | 11.250 |
| Q6_K | 178 / 320 | 11.125 |
| BF16 | 171 / 320 | 10.688 |
| Q4_K_M | 162 / 320 | 10.125 |
The evaluator did not receive model identities until the score set had been frozen.
The evaluation used a 512-token generation ceiling. The evaluator noted truncation or abrupt cutoff in 65 of 80 responses. Treat this as a controlled quantization-retention check, not as a complete Building Engineering benchmark.
Conversion
The GGUF artifacts were produced and runtime-qualified with llama.cpp.
The BF16 GGUF is the canonical parent of the Q8_0, Q6_K, Q5_K_M, and Q4_K_M files.
Full model
Transformers/BF16 release:
Irfanuruchi/OLMo-2-1124-7B-BuildEng
Intended use
Research, education, local inference, technical drafting, and Building Engineering experiments.
This model does not replace project-specific engineering analysis, code review, site investigation, or professional judgment.
License
Apache-2.0.
See the full-model repository and upstream OLMo 2 documentation for provenance and responsible-use information.
Run Irfanuruchi/OLMo-2-1124-7B-BuildEng-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models