How to use from
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 Forris/learnlanguage-models
# Run inference directly in the terminal:
llama cli -hf Forris/learnlanguage-models
Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Forris/learnlanguage-models
# Run inference directly in the terminal:
llama cli -hf Forris/learnlanguage-models
Build from source code
git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf Forris/learnlanguage-models
# Run inference directly in the terminal:
./build/bin/llama cli -hf Forris/learnlanguage-models
Use Docker
docker model run hf.co/Forris/learnlanguage-models
Quick Links

LearnLanguage models

Models the LearnLanguage app (a language-learning app) downloads to run on the phone. Each file is pinned by the app to a commit of this repo. Licenses differ per file:

grader.gguf β€” answer checker

GGUF (Q4_K_M) of Qwen/Qwen3-1.7B fine-tuned with LoRA (merged) to check whether a learner's free answer does what an exercise asks. Answers with a language code and one 1/0 per item.

  • License: Apache 2.0, as the base model.
  • Training data: generated by the app's own scripts (templates written for it), no third-party data.

whisper.bin β€” Serbian speech recognition

whisper.cpp model (ggml, q5_0) of openai/whisper-large-v3-turbo, fine-tuned for Serbian with LoRA (merged): transcripts in Latin script with case and punctuation. Word error rate (whisper.cpp, greedy): Common Voice sr, held-out speakers β€” 9.4% (base model 39.8%); FLEURS sr_rs test β€” 12.1% (base model 13.3%).

  • License: MIT, as the base model.
  • Training data: FLEURS sr_rs (CC BY 4.0, Google) and Common Voice 27.0 sr (CC0, Mozilla).

partner.gguf β€” conversation partner

GGUF (Q4_K_M) of Gemma 4 E4B by Google, unchanged.

  • License: Apache 2.0, as the original model.
Downloads last month
-
GGUF
Model size
2B params
Architecture
qwen3
Hardware compatibility
Log In to add your hardware

We're not able to determine the quantization variants.

Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support