Instructions to use Rohanify/PyBlissa-Coder-50M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Rohanify/PyBlissa-Coder-50M with Transformers:

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

pipe = pipeline("text-generation", model="Rohanify/PyBlissa-Coder-50M")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Rohanify/PyBlissa-Coder-50M", dtype="auto")

llama-cpp-python

How to use Rohanify/PyBlissa-Coder-50M with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Rohanify/PyBlissa-Coder-50M",
	filename="PyBlissa-Coder-50M-F32.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Rohanify/PyBlissa-Coder-50M 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 Rohanify/PyBlissa-Coder-50M:F32
# Run inference directly in the terminal:
llama cli -hf Rohanify/PyBlissa-Coder-50M:F32

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Rohanify/PyBlissa-Coder-50M:F32
# Run inference directly in the terminal:
llama cli -hf Rohanify/PyBlissa-Coder-50M:F32

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 Rohanify/PyBlissa-Coder-50M:F32
# Run inference directly in the terminal:
./llama-cli -hf Rohanify/PyBlissa-Coder-50M:F32

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 Rohanify/PyBlissa-Coder-50M:F32
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Rohanify/PyBlissa-Coder-50M:F32

Use Docker

docker model run hf.co/Rohanify/PyBlissa-Coder-50M:F32

LM Studio
Jan

vLLM

How to use Rohanify/PyBlissa-Coder-50M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Rohanify/PyBlissa-Coder-50M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Rohanify/PyBlissa-Coder-50M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Rohanify/PyBlissa-Coder-50M:F32

SGLang

How to use Rohanify/PyBlissa-Coder-50M 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 "Rohanify/PyBlissa-Coder-50M" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Rohanify/PyBlissa-Coder-50M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Rohanify/PyBlissa-Coder-50M" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Rohanify/PyBlissa-Coder-50M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use Rohanify/PyBlissa-Coder-50M with Ollama:
```
ollama run hf.co/Rohanify/PyBlissa-Coder-50M:F32
```

Unsloth Studio

How to use Rohanify/PyBlissa-Coder-50M 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 Rohanify/PyBlissa-Coder-50M 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 Rohanify/PyBlissa-Coder-50M to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Rohanify/PyBlissa-Coder-50M to start chatting

Atomic Chat new
Docker Model Runner
How to use Rohanify/PyBlissa-Coder-50M with Docker Model Runner:
```
docker model run hf.co/Rohanify/PyBlissa-Coder-50M:F32
```

Lemonade

How to use Rohanify/PyBlissa-Coder-50M with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Rohanify/PyBlissa-Coder-50M:F32

Run and chat with the model

lemonade run user.PyBlissa-Coder-50M-F32

List all available models

lemonade list

PyBlissa-Coder-50M / README.md

Rohanify

Update README.md

7f96631 verified 6 days ago

preview code

Raw

History Blame Contribute Delete

3.26 kB

	---
	license: mit
	language:
	- en
	library_name: transformers
	pipeline_tag: text-generation
	tags:
	- code
	- python
	- gguf
	- small-model
	- pretrained-from-scratch
	- gpt2
	- from-scratch
	- coding
	- SLM
	---

	# 🚀 PyBlissa-Coder-50M
	# !! BENCHMARK ON HumaEval DATASET: 10.4% !!

	A 50M-parameter Python code generation model trained from scratch on a single RTX 5080. Built as part of the PRIME lineup of small, locally-runnable AI systems.

	Despite its size, PyBlissa punches well above its weight on Python instruction-following tasks. Trained near-Chinchilla optimal (~13 tokens/parameter) for maximum capacity utilization.

	This model punched a solid 10.4% score in OpenAI's HumanEval dataset, which is an amazing number for this model's size!
	However, this model technically can generate bad outputs. You'd need to tweak the temperature. But that's a rare case!

	![Benchmark](pyblissa_humaneval_benchmark.png)


	![Loss Visualization](loss_visualization.png)



	## Stats

	\| \| \|
	\|---\|---\|
	\| Parameters \| 50.2M \|
	\| Architecture \| Decoder-only transformer (GPT-2 style) \|
	\| Context length \| 1024 tokens \|
	\| Vocab size \| 16,000 (custom ByteLevel BPE) \|
	\| Train tokens \| 166M \|
	\| Final val loss \| 0.474 \|
	\| Training time \| 73 minutes (RTX 5080) \|

	## Architecture

	```
	d_model: 640
	n_layer: 8
	n_head: 8
	d_ff: 2560
	block_size: 1024
	tied embeddings, pre-LN, no bias, GELU MLP, SDPA attention
	```

	## Training data

	Two-source code-instruction corpus, 425k samples → 166M tokens after BPE tokenization:

	- `nvidia/OpenCodeInstruct` — 400k high-quality instruction-code pairs
	- `flytech/python-codes-25k` — 25k Python-focused instruction-code pairs

	Trained for 4 epochs with cosine LR schedule (3e-4 → 3e-5), bf16 autocast, batch size 20.

	## Prompt format

	Trained on a strict prefix structure:

	```
	PROMPT: <your instruction>
	CODE:
	<generated code>
	```

	Anything else is out-of-distribution. The Modelfile in this repo handles the formatting automatically.

	## Usage — Ollama (recommended)

	```bash
	ollama run hf.co/Rohanify/PyBlissa-Coder-50M:F32
	```

	Or pull the GGUF directly and create locally:

	```bash
	ollama create pyblissa-coder -f Modelfile
	ollama run pyblissa-coder "write a function to merge two sorted lists"
	```

	## Usage — llama.cpp

	```bash
	./llama-cli -m PyBlissa-Coder-50M-F32.gguf \
	-p "PROMPT: write a function to reverse a string\nCODE:\n" \
	--temp 0.8 --top-k 50 --top-p 0.95 -n 256
	```

	## Files

	\| File \| Purpose \|
	\|---\|---\|
	\| `PyBlissa-Coder-50M-F32.gguf` \| Full-precision GGUF weights \|
	\| `Modelfile` \| Ollama config (prompt template, stop tokens, sampling) \|
	\| `tokenizer.json` \| Custom 16k BPE tokenizer \|

	## Limitations

	- Mostly Python — other languages weren't that much in training data
	- 1024-token context — longer programs get truncated
	- Small flytech subset (~6% of training data) contains code with unescaped quote bugs; the model occasionally inherits this pattern
	- No safety tuning, no RLHF — base model only

	## Acknowledgments

	Datasets by NVIDIA and flytech. Built using the nanoGPT-style training recipe with custom tokenization. Tooling: PyTorch, HuggingFace `tokenizers`, llama.cpp for GGUF conversion.

	---

	Made by Rohan. Also known as ElectroPlayin on YouTube