Instructions to use SupraLabs/Supra-50M-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use SupraLabs/Supra-50M-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="SupraLabs/Supra-50M-Instruct")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("SupraLabs/Supra-50M-Instruct")
model = AutoModelForCausalLM.from_pretrained("SupraLabs/Supra-50M-Instruct")

llama-cpp-python

How to use SupraLabs/Supra-50M-Instruct with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="SupraLabs/Supra-50M-Instruct",
	filename="Supra-50M-f16.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 SupraLabs/Supra-50M-Instruct 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 SupraLabs/Supra-50M-Instruct:F16
# Run inference directly in the terminal:
llama cli -hf SupraLabs/Supra-50M-Instruct:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf SupraLabs/Supra-50M-Instruct:F16
# Run inference directly in the terminal:
llama cli -hf SupraLabs/Supra-50M-Instruct: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 SupraLabs/Supra-50M-Instruct:F16
# Run inference directly in the terminal:
./llama-cli -hf SupraLabs/Supra-50M-Instruct: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 SupraLabs/Supra-50M-Instruct:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf SupraLabs/Supra-50M-Instruct:F16

Use Docker

docker model run hf.co/SupraLabs/Supra-50M-Instruct:F16

LM Studio
Jan

vLLM

How to use SupraLabs/Supra-50M-Instruct with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/SupraLabs/Supra-50M-Instruct:F16

SGLang

How to use SupraLabs/Supra-50M-Instruct 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 "SupraLabs/Supra-50M-Instruct" \
    --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": "SupraLabs/Supra-50M-Instruct",
		"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 "SupraLabs/Supra-50M-Instruct" \
        --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": "SupraLabs/Supra-50M-Instruct",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use SupraLabs/Supra-50M-Instruct with Ollama:
```
ollama run hf.co/SupraLabs/Supra-50M-Instruct:F16
```

Unsloth Studio

How to use SupraLabs/Supra-50M-Instruct 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 SupraLabs/Supra-50M-Instruct 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 SupraLabs/Supra-50M-Instruct to start chatting

Using HuggingFace Spaces for Unsloth

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

Atomic Chat new
Docker Model Runner
How to use SupraLabs/Supra-50M-Instruct with Docker Model Runner:
```
docker model run hf.co/SupraLabs/Supra-50M-Instruct:F16
```

Lemonade

How to use SupraLabs/Supra-50M-Instruct with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull SupraLabs/Supra-50M-Instruct:F16

Run and chat with the model

lemonade run user.Supra-50M-Instruct-F16

List all available models

lemonade list

SupraLabs/Supra-50M-Instruct

by aizexintong - opened May 28

base: refs/heads/main

←

from: refs/pr/4

Discussion Files changed

-27

Files changed (3) hide show

.gitattributes +0 -1
README.md +2 -23
Supra-50M-f16.gguf → supra-50m-instruct-f16.gguf +2 -2

.gitattributes CHANGED Viewed

@@ -34,4 +34,3 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 supra-50m-instruct-f16.gguf filter=lfs diff=lfs merge=lfs -text
-Supra-50M-f16.gguf filter=lfs diff=lfs merge=lfs -text

 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 supra-50m-instruct-f16.gguf filter=lfs diff=lfs merge=lfs -text

README.md CHANGED Viewed

@@ -20,7 +20,6 @@ tags:
 - slm
 base_model:
 - SupraLabs/Supra-50M-Base
-new_version: SupraLabs/Supra-1.5-50M-Instruct-exp
 ---
 # 🦅 Supra-50M Instruct
@@ -78,10 +77,8 @@ For more details, the full code, configs and weights, please refer to [https://h
 ```python
 import os
 import warnings
-import time
 os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
-os.environ["CUDA_VISIBLE_DEVICES"] = "0"
 warnings.filterwarnings("ignore", category=UserWarning, module="transformers")
 import torch
@@ -89,13 +86,6 @@ from transformers import pipeline, AutoTokenizer, logging
 logging.set_verbosity_error()
-# ── Global variables ──────────────────────────────────────────────────────────
-end = time.time()
-start = time.time()
-tokens = []
 # ── Config ────────────────────────────────────────────────────────────────────
 MODEL_ID = "SupraLabs/Supra-50M-Instruct"
@@ -140,7 +130,7 @@ def build_prompt(instruction: str, input_text: str = "") -> str:
 def generate(instruction: str, input_text: str = "", max_new_tokens: int = MAX_NEW_TOKENS) -> str:
     prompt = build_prompt(instruction, input_text)
-    start = time.time()
     result = pipe(
         prompt,
         max_new_tokens=max_new_tokens,
@@ -154,13 +144,7 @@ def generate(instruction: str, input_text: str = "", max_new_tokens: int = MAX_N
         return_full_text=False,
         generation_config=None
     )
-    end = time.time()
-    generated_text = result[0]["generated_text"]
-    tokens = pipe.tokenizer(generated_text)["input_ids"]
-    return generated_text, tokens, end, start
 # ── Interactive loop ──────────────────────────────────────────────────────────
 if __name__ == "__main__":
@@ -174,13 +158,8 @@ if __name__ == "__main__":
         inp = input("Input (optional, Enter to skip): ").strip()
         print("-" * 50)
-        text, tokens, end, start = generate(instruction, inp)
-        print(text)
         print()
-        print(f"Generated tokens: {len(tokens)}")
-        print(f"Time: {end - start:.2f}s")
-        print(f"Speed: {len(tokens) / (end - start):.2f} tokens/sec")
 ```
 ---

 - slm
 base_model:
 - SupraLabs/Supra-50M-Base
 ---
 # 🦅 Supra-50M Instruct
 ```python
 import os
 import warnings
 os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
 warnings.filterwarnings("ignore", category=UserWarning, module="transformers")
 import torch
 logging.set_verbosity_error()
 # ── Config ────────────────────────────────────────────────────────────────────
 MODEL_ID = "SupraLabs/Supra-50M-Instruct"
 def generate(instruction: str, input_text: str = "", max_new_tokens: int = MAX_NEW_TOKENS) -> str:
     prompt = build_prompt(instruction, input_text)
     result = pipe(
         prompt,
         max_new_tokens=max_new_tokens,
         return_full_text=False,
         generation_config=None
     )
 # ── Interactive loop ──────────────────────────────────────────────────────────
 if __name__ == "__main__":
         inp = input("Input (optional, Enter to skip): ").strip()
         print("-" * 50)
+        print(generate(instruction, inp))
         print()
 ```
 ---

Supra-50M-f16.gguf → supra-50m-instruct-f16.gguf RENAMED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:14a5840b3e66cd159eae57675dab4ec4e8eb074b823ee68c2f65eb58a655f6f7
-size 104739616

 version https://git-lfs.github.com/spec/v1
+oid sha256:e104f91e7aeed5c11d1601461b93ddc6b71c8d607d064c3afd8d0bfe5590647f
+size 104741056