Instructions to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF",
	filename="TinyLlama-1.1B-RYS-10-14-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

Use Docker

docker model run hf.co/john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with Ollama:
```
ollama run hf.co/john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
```

Unsloth Studio

How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF to start chatting

Atomic Chat new
Docker Model Runner
How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with Docker Model Runner:
```
docker model run hf.co/john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M
```

Lemonade

How to use john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.TinyLlama-1.1B-RYS-10-14-GGUF-Q4_K_M

List all available models

lemonade list

TinyLlama-1.1B-RYS-10-14

TinyLlama-1.1B-Chat-v1.0 with layers 10-13 duplicated. The mid-stack block runs twice — and it is the EQ circuit, not reasoning, that recovers most.

22 base layers → 26 after duplication. No training, no merging, no weight changes.

EQ 4.65 → 52.50 (+47.85). Math 0.296 → 0.2964 (+0.04). Reasoning 29.41% → 23.53% (−5.88).

Results

Metric	Baseline	RYS (10,14)	Delta
Math	0.296	0.2964	+0.04
EQ	4.65	52.50	+47.85
Reasoning	29.41%	23.53%	−5.88

The EQ unlock. TinyLlama-1.1B has the lowest baseline EQ in the v2 corpus by an order of magnitude (4.65, where the next-lowest is the 27.11 of Llama-3.2-1B). RYS recovers it dramatically: +47.85 EQ on this single configuration, the largest EQ lift anywhere in the 21-model corpus.

This is the data point that refined the corpus-wide hypothesis from "weak baselines lift more (in reasoning)" to "weak baselines lift more in the dimension where they're weakest." Most v2 models lift reasoning because reasoning is their weakest probe; TinyLlama is the case where it isn't — EQ is — and the recovery follows the weakness.

Pick this when you want a tiny model whose conversational fluency is the goal. Reasoning is unchanged-to-slightly-worse; math is flat. EQ is the only dimension that meaningfully moves at this config.

Usage

llama-server -m TinyLlama-1.1B-RYS-10-14-Q4_K_M.gguf -ngl 99

Full sweep data

26 configurations tested. (10,14) block-4 is the best-combined pick (peak EQ Δ + minimal cost elsewhere). Full per-config sweep + cross-architecture analysis: v2 dataset.

Part of the RYS Sovereign Collection v2.

Where this sits in the Sovereign Collection

v1 — Qwen2.5 cross-scale + Qwen3-32B headline crossover. 5 model repos.

v2 — cross-architecture corpus. 21 model variants across 10 architecture families. Inverse correlation (r = −0.726): weak baselines lift more, in their weakest dimension. Three reasoning-recovery mechanisms; this model is the dimension-specific weakness-recovery exemplar that refined the hypothesis. 13 deployable RYS-applied weight repos covering every non-zero-lift variant.

Credit

John Broadway, with collaboration from Claude (Opus 4.6 in April 2026 sweep generation and build pipeline; Opus 4.7 in May 2026 cross-architecture analysis and publication). Original RYS method by David Ng on Qwen2-72B; sweep + probe toolkit by alainnothere.

Downloads last month: 164

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF

Base model

TinyLlama/TinyLlama-1.1B-Chat-v1.0

Quantized

(149)

this model

Collection including john-broadway/TinyLlama-1.1B-RYS-10-14-GGUF

RYS Sovereign Collection v2

Collection

Qwen3-family cohort (0.6B/1.7B/8B) + cross-architecture sweep corpus (21 variants, 10 families). v1 collection: Qwen2.5 cross-scale + Qwen3-32B. • 18 items • Updated 29 days ago