Instructions to use louhless/Ycoder-medium with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use louhless/Ycoder-medium with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="louhless/Ycoder-medium")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("louhless/Ycoder-medium", dtype="auto")

llama-cpp-python

How to use louhless/Ycoder-medium with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="louhless/Ycoder-medium",
	filename="Ycoder-medium-f16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use louhless/Ycoder-medium with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf louhless/Ycoder-medium:F16
# Run inference directly in the terminal:
llama-cli -hf louhless/Ycoder-medium:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf louhless/Ycoder-medium:F16
# Run inference directly in the terminal:
llama-cli -hf louhless/Ycoder-medium:F16

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 louhless/Ycoder-medium:F16
# Run inference directly in the terminal:
./llama-cli -hf louhless/Ycoder-medium:F16

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 louhless/Ycoder-medium:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf louhless/Ycoder-medium:F16

Use Docker

docker model run hf.co/louhless/Ycoder-medium:F16

LM Studio
Jan

vLLM

How to use louhless/Ycoder-medium with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/louhless/Ycoder-medium:F16

SGLang

How to use louhless/Ycoder-medium with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "louhless/Ycoder-medium" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "louhless/Ycoder-medium",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "louhless/Ycoder-medium" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "louhless/Ycoder-medium",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use louhless/Ycoder-medium with Ollama:
```
ollama run hf.co/louhless/Ycoder-medium:F16
```

Unsloth Studio

How to use louhless/Ycoder-medium 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 louhless/Ycoder-medium 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 louhless/Ycoder-medium to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for louhless/Ycoder-medium to start chatting

Docker Model Runner
How to use louhless/Ycoder-medium with Docker Model Runner:
```
docker model run hf.co/louhless/Ycoder-medium:F16
```

Lemonade

How to use louhless/Ycoder-medium with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull louhless/Ycoder-medium:F16

Run and chat with the model

lemonade run user.Ycoder-medium-F16

List all available models

lemonade list

louhless commited on 10 days ago

Commit

703e17a

verified ·

1 Parent(s): aa9504b

Update README.md

Browse files

Files changed (1) hide show

README.md +97 -0

README.md CHANGED Viewed

@@ -1,3 +1,100 @@
 ---
 license: apache-2.0
 ---

 ---
+language:
+- en
+- de
+tags:
+- text-generation
+- causal-lm
+- fine-tuned
+- gguf
+- code
+- python
+- glsl
+- opengl
+- german
+- news
+- experimental
 license: apache-2.0
+base_model: Qwen/Qwen2.5-Coder-0.5B-Instruct
+pipeline_tag: text-generation
+library_name: transformers
+model_creator: louhless
 ---
+# Ycoder-medium
+`Ycoder-medium` is an experimental local fine-tune of `Qwen/Qwen2.5-Coder-0.5B-Instruct` created by **louhless**.
+It is targeted at:
+- OpenGL / GLSL
+- Python
+- German replies
+- cautious 2025-2026 news and public-health summaries
+## Important Note
+This model is **not trained from scratch**.
+It is a small LoRA fine-tune on top of `Qwen/Qwen2.5-Coder-0.5B-Instruct`.
+The goal is to improve behavior in a narrow target set. Any “15% improvement” claim should be treated as a target, not a verified benchmark result, unless evaluated on a fixed benchmark before and after training.
+## Model Details
+- **Model name:** `Ycoder-medium`
+- **Creator:** `louhless`
+- **Base model:** `Qwen/Qwen2.5-Coder-0.5B-Instruct`
+- **Architecture:** Qwen2 causal language model
+- **Context length:** 32768
+- **Language:** English and German
+- **Export:** GGUF available
+- **Status:** experimental
+## Training Focus
+The model was tuned for:
+- Python utility code
+- Python code explanations
+- GLSL fragment shaders
+- GLSL vertex shaders
+- OpenGL concepts such as VAO/VBO
+- German short-form answers
+- simple math
+- cautious dated summaries for 2025-2026 public-health/news topics
+## News / Health Safety
+For topics such as Hantavirus, the project uses both small fine-tuning examples and local dated context snippets.
+This is intentional: recent news and public-health information should not be trusted from model weights alone.
+The model should:
+- answer cautiously
+- mention dates when relevant
+- avoid medical diagnosis
+- avoid treatment promises
+- recommend official sources such as WHO, CDC, ECDC, or local health authorities
+It should **not** be used for diagnosis or medical decision-making.
+## Training Data
+The initial custom dataset includes examples for:
+- Python utility functions and explanations
+- GLSL shaders and OpenGL concepts
+- German short answers
+- simple math
+- dated 2025-2026 Hantavirus summaries based on WHO, CDC, and ECDC public information
+## Example Prompts
+### Python
+Prompt:
+```text
+Write Python code to read a JSON file safely.