Instructions to use sol-wy/talkie-1930-13b-it-q5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use sol-wy/talkie-1930-13b-it-q5 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf sol-wy/talkie-1930-13b-it-q5
# Run inference directly in the terminal:
llama cli -hf sol-wy/talkie-1930-13b-it-q5

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf sol-wy/talkie-1930-13b-it-q5
# Run inference directly in the terminal:
llama cli -hf sol-wy/talkie-1930-13b-it-q5

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 sol-wy/talkie-1930-13b-it-q5
# Run inference directly in the terminal:
./llama-cli -hf sol-wy/talkie-1930-13b-it-q5

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 sol-wy/talkie-1930-13b-it-q5
# Run inference directly in the terminal:
./build/bin/llama-cli -hf sol-wy/talkie-1930-13b-it-q5

Use Docker

docker model run hf.co/sol-wy/talkie-1930-13b-it-q5

LM Studio
Jan

vLLM

How to use sol-wy/talkie-1930-13b-it-q5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sol-wy/talkie-1930-13b-it-q5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sol-wy/talkie-1930-13b-it-q5",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sol-wy/talkie-1930-13b-it-q5

Ollama
How to use sol-wy/talkie-1930-13b-it-q5 with Ollama:
```
ollama run hf.co/sol-wy/talkie-1930-13b-it-q5
```
Unsloth Desktop
Docker Model Runner
How to use sol-wy/talkie-1930-13b-it-q5 with Docker Model Runner:
```
docker model run hf.co/sol-wy/talkie-1930-13b-it-q5
```

Lemonade

How to use sol-wy/talkie-1930-13b-it-q5 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull sol-wy/talkie-1930-13b-it-q5

Run and chat with the model

lemonade run user.talkie-1930-13b-it-q5-{{QUANT_TAG}}

List all available models

lemonade list

Atomic Chat

Talkie 1930 13B IT Q5

talkie-1930-13b-it-q5 is an experimental Q5_0 GGUF build derived from talkie-lm/talkie-1930-13b-it.

This is a consumer-GPU-oriented artifact for running Talkie locally through a patched llama.cpp runtime. It is not yet compatible with stock llama.cpp or stock LM Studio because Talkie uses a custom architecture.

File

talkie-1930-13b-it-q5.gguf

Expected SHA256:

B6025276018B228CB35CDC76F2D957EB8037EA595F7C9486BC2971ECAFEAD0BA

Runtime

Use the companion GitHub repo for the Node app, patched llama.cpp runtime, build scripts, and release notes:

https://github.com/solwyc/talkie-1930-13b-it-q5

Tested local Windows settings:

GPU:          RTX 4070 12 GB
GGUF:         Q5_0, about 9.13 GB
offload:      -ngl 26
context:      1024
speed:        roughly 7-8 tok/s in short local tests

The repo includes experimental Linux CUDA and macOS Metal build scaffolding, but those ports were drafted blindly with GPT-5.5 assistance and have not been maintainer-tested yet.

Important Compatibility Note

This GGUF has:

general.architecture = talkie

It requires a Talkie-aware llama.cpp runtime with support for Talkie's custom blocks, tensor mapping, QK norm/gain tensors, and RMSNorm epsilon behavior.