Instructions to use prithivMLmods/amos-1-medium-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prithivMLmods/amos-1-medium-GGUF with Transformers:

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

pipe = pipeline("image-text-to-text", model="prithivMLmods/amos-1-medium-GGUF")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("prithivMLmods/amos-1-medium-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use prithivMLmods/amos-1-medium-GGUF 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 prithivMLmods/amos-1-medium-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

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 prithivMLmods/amos-1-medium-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

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 prithivMLmods/amos-1-medium-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Use Docker

docker model run hf.co/prithivMLmods/amos-1-medium-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use prithivMLmods/amos-1-medium-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prithivMLmods/amos-1-medium-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prithivMLmods/amos-1-medium-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/prithivMLmods/amos-1-medium-GGUF:Q4_K_M

SGLang

How to use prithivMLmods/amos-1-medium-GGUF 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 "prithivMLmods/amos-1-medium-GGUF" \
    --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": "prithivMLmods/amos-1-medium-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "prithivMLmods/amos-1-medium-GGUF" \
        --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": "prithivMLmods/amos-1-medium-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Ollama
How to use prithivMLmods/amos-1-medium-GGUF with Ollama:
```
ollama run hf.co/prithivMLmods/amos-1-medium-GGUF:Q4_K_M
```
Unsloth Desktop

How to use prithivMLmods/amos-1-medium-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "prithivMLmods/amos-1-medium-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use prithivMLmods/amos-1-medium-GGUF with Docker Model Runner:
```
docker model run hf.co/prithivMLmods/amos-1-medium-GGUF:Q4_K_M
```

Lemonade

How to use prithivMLmods/amos-1-medium-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.amos-1-medium-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use prithivMLmods/amos-1-medium-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use prithivMLmods/amos-1-medium-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf prithivMLmods/amos-1-medium-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "prithivMLmods/amos-1-medium-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

amos-1-medium-GGUF

Amos 1 medium is a 9B decision model built on Qwen3.5-9B, the middle of the four-size Amos 1 family (large, medium, small, mini). Given a case as text, JSON or a photo, plus a schema of typed questions, it answers every question in one forward pass with a probability for each allowed answer, and it writes no text, so there is nothing to parse. Question types are yes/no (noul), pick one (choice, up to 255 options), rate (score), pick any (multi), quote (span) and point (region, with a photo). On the typed-decisions test split (2,000 decisions from four unseen workflows, zero-shot) it scores 76.5% accuracy, KL 0.151 and Brier 0.080. That is ahead of Perplexity Decider v1.1 (73.9%) and Cloudflare's Clef (72.5%) and just under Decider 1 (76.8%, its makers' own figure) and Microsoft-Decision-1 (79.5%). The card says it is level with the 27B large at half its server time. The author ran the rival numbers except Decider 1's. With per-question priors fitted on the benchmark's train split, it gets 76.7% against Decider 1's 80.1%. On Decision Bench it scores 89.1%, below the leaderboard's 93.8% to 94.5% for Jev 1.13 and Perplexity Decider v1.1. Its Decision Index 0.3 public score is 52.55, run with the board's own kit but without the private tests the board's ranking adds. JevBench public items come out at 100.0 / 98.6 / 73.0 (easy / standard / hard), and median server time is 61 ms for a three-question request. It runs on a 24 GB GPU and reads up to 131,072 tokens (about 40 GB near that length). It is available through a hosted keyless API or a self-hosted amos.serve package. The listed limits are that it is behind the large on multi-step or broad-knowledge questions, rating answers are the least reliable, and it is English only. It is released under Apache 2.0.

Amos 1 Medium model available @ — https://huggingface.co/mukaddamzaid/amos-1-medium/tree/main/model

MTP — Speculative Decoding: No

Model Files

File Name	Quant Type	File Size	File Link	Description
amos-1-medium.BF16.gguf	BF16	17.9 GB	Link	Full BF16 weights. Highest quality, largest file size.
amos-1-medium.Q3_K_L.gguf	Q3_K_L	4.93 GB	Link	Lower quality but usable, good for low RAM availability.
amos-1-medium.Q3_K_M.gguf	Q3_K_M	4.62 GB	Link	Low quality.
amos-1-medium.Q4_K_M.gguf	Q4_K_M	5.63 GB	Link	Good quality, default size for most use cases, recommended.
amos-1-medium.Q4_K_S.gguf	Q4_K_S	5.35 GB	Link	Slightly lower quality with more space savings, recommended.
amos-1-medium.Q5_K_M.gguf	Q5_K_M	6.47 GB	Link	High quality, recommended.
amos-1-medium.Q5_K_S.gguf	Q5_K_S	6.31 GB	Link	High quality, recommended.
amos-1-medium.Q6_K.gguf	Q6_K	7.36 GB	Link	Very high quality, near perfect, recommended.
amos-1-medium.mmproj-bf16.gguf	mmproj-bf16	922 MB	Link	Multimodal projection file in BF16 format. Used for vision/language models.