Instructions to use gbuzhf/Sage-A1.5-Sangreal-ICE with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use gbuzhf/Sage-A1.5-Sangreal-ICE 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 gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0
# Run inference directly in the terminal:
llama cli -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0
# Run inference directly in the terminal:
llama cli -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

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 gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0
# Run inference directly in the terminal:
./build/bin/llama cli -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Use Docker

docker model run hf.co/gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Llama new
LM Studio
Jan

vLLM

How to use gbuzhf/Sage-A1.5-Sangreal-ICE with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gbuzhf/Sage-A1.5-Sangreal-ICE"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gbuzhf/Sage-A1.5-Sangreal-ICE",
		"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/gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Ollama
How to use gbuzhf/Sage-A1.5-Sangreal-ICE with Ollama:
```
ollama run hf.co/gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0
```
Unsloth Desktop

How to use gbuzhf/Sage-A1.5-Sangreal-ICE with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

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": "gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use gbuzhf/Sage-A1.5-Sangreal-ICE with Docker Model Runner:
```
docker model run hf.co/gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0
```

Lemonade

How to use gbuzhf/Sage-A1.5-Sangreal-ICE with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Run and chat with the model

lemonade run user.Sage-A1.5-Sangreal-ICE-Q8_0

List all available models

lemonade list

Hermes Agent

How to use gbuzhf/Sage-A1.5-Sangreal-ICE with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

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 gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use gbuzhf/Sage-A1.5-Sangreal-ICE with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0

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 "gbuzhf/Sage-A1.5-Sangreal-ICE:Q8_0" \
  --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"

Sage-A1.5-Sangreal-ICE / assets /code16k_kld_vs_size.svg

gbuzhf

chart

3a688de verified 7 days ago

preview code

Raw History Blame Contribute Delete

5.12 kB

<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 860 540" width="860" height="540" font-family="Inter,Segoe UI,Helvetica,Arial,sans-serif">
<rect width="860" height="540" fill="#fcfcfb"/>
<text x="84" y="30" font-size="19" font-weight="600" fill="#0b0b0b">Sage-A1.5-Sangreal-ICE: KL divergence vs size (code, 16K context)</text>
<text x="84" y="52" font-size="13" fill="#52514e">Mean KLD vs Sage-A1.5 F16 on the same tokens, llama-perplexity, n_ctx 16384, 8 chunks. Lower-left is better.</text>
<line x1="84" x2="830" y1="470.0" y2="470.0" stroke="#e8e7e3"/>
<text x="76" y="474.0" font-size="12" fill="#898781" text-anchor="end">0.000</text>
<line x1="84" x2="830" y1="395.2" y2="395.2" stroke="#e8e7e3"/>
<text x="76" y="399.2" font-size="12" fill="#898781" text-anchor="end">0.025</text>
<line x1="84" x2="830" y1="320.4" y2="320.4" stroke="#e8e7e3"/>
<text x="76" y="324.4" font-size="12" fill="#898781" text-anchor="end">0.050</text>
<line x1="84" x2="830" y1="245.6" y2="245.6" stroke="#e8e7e3"/>
<text x="76" y="249.6" font-size="12" fill="#898781" text-anchor="end">0.075</text>
<line x1="84" x2="830" y1="170.8" y2="170.8" stroke="#e8e7e3"/>
<text x="76" y="174.8" font-size="12" fill="#898781" text-anchor="end">0.100</text>
<line x1="84" x2="830" y1="96.0" y2="96.0" stroke="#e8e7e3"/>
<text x="76" y="100.0" font-size="12" fill="#898781" text-anchor="end">0.125</text>
<line x1="84.0" x2="84.0" y1="470" y2="475" stroke="#898781"/>
<text x="84.0" y="490" font-size="12" fill="#898781" text-anchor="middle">10</text>
<line x1="177.2" x2="177.2" y1="470" y2="475" stroke="#898781"/>
<text x="177.2" y="490" font-size="12" fill="#898781" text-anchor="middle">12</text>
<line x1="270.5" x2="270.5" y1="470" y2="475" stroke="#898781"/>
<text x="270.5" y="490" font-size="12" fill="#898781" text-anchor="middle">14</text>
<line x1="363.8" x2="363.8" y1="470" y2="475" stroke="#898781"/>
<text x="363.8" y="490" font-size="12" fill="#898781" text-anchor="middle">16</text>
<line x1="457.0" x2="457.0" y1="470" y2="475" stroke="#898781"/>
<text x="457.0" y="490" font-size="12" fill="#898781" text-anchor="middle">18</text>
<line x1="550.2" x2="550.2" y1="470" y2="475" stroke="#898781"/>
<text x="550.2" y="490" font-size="12" fill="#898781" text-anchor="middle">20</text>
<line x1="643.5" x2="643.5" y1="470" y2="475" stroke="#898781"/>
<text x="643.5" y="490" font-size="12" fill="#898781" text-anchor="middle">22</text>
<line x1="736.8" x2="736.8" y1="470" y2="475" stroke="#898781"/>
<text x="736.8" y="490" font-size="12" fill="#898781" text-anchor="middle">24</text>
<line x1="830.0" x2="830.0" y1="470" y2="475" stroke="#898781"/>
<text x="830.0" y="490" font-size="12" fill="#898781" text-anchor="middle">26</text>
<line x1="84" x2="830" y1="470" y2="470" stroke="#898781"/>
<text x="457.0" y="518" font-size="13" fill="#52514e" text-anchor="middle">Model size without the MTP draft head (GB, lower is better)</text>
<text transform="translate(24,283.0) rotate(-90)" font-size="13" fill="#52514e" text-anchor="middle">Mean KL divergence (lower is better)</text>
<path d="M172.4,107.0 L294.5,249.1 L479.0,366.0 L667.6,387.6 L761.5,395.1" fill="none" stroke="#1baf7a" stroke-width="2"/>
<line x1="172.4" x2="172.4" y1="115.2" y2="98.8" stroke="#1baf7a" stroke-width="1.5"/>
<polygon points="172.4,100.0 178.9,112.0 165.9,112.0" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
<text x="182.4" y="98.0" font-size="12" fill="#0b0b0b" text-anchor="start">12G-ICEv2 <tspan fill="#52514e">0.1213</tspan></text>
<line x1="294.5" x2="294.5" y1="255.5" y2="242.6" stroke="#1baf7a" stroke-width="1.5"/>
<polygon points="294.5,242.1 301.0,254.1 288.0,254.1" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
<text x="304.5" y="240.1" font-size="12" fill="#0b0b0b" text-anchor="start">15G-ICEv2 <tspan fill="#52514e">0.0738</tspan></text>
<line x1="479.0" x2="479.0" y1="370.5" y2="361.6" stroke="#1baf7a" stroke-width="1.5"/>
<polygon points="479.0,359.0 485.5,371.0 472.5,371.0" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
<text x="489.0" y="357.0" font-size="12" fill="#0b0b0b" text-anchor="start">19G-ICEv2 <tspan fill="#52514e">0.0347</tspan></text>
<line x1="667.6" x2="667.6" y1="391.9" y2="383.3" stroke="#1baf7a" stroke-width="1.5"/>
<polygon points="667.6,380.6 674.1,392.6 661.1,392.6" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
<text x="677.6" y="378.6" font-size="12" fill="#0b0b0b" text-anchor="start">23G-ICEv2 <tspan fill="#52514e">0.0275</tspan></text>
<line x1="761.5" x2="761.5" y1="399.3" y2="390.9" stroke="#1baf7a" stroke-width="1.5"/>
<polygon points="761.5,388.1 768.0,400.1 755.0,400.1" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
<text x="761.5" y="417.1" font-size="12" fill="#0b0b0b" text-anchor="middle">25G-ICEv2 <tspan fill="#52514e">0.0250</tspan></text>
<polygon points="90,72 96.5,84 83.5,84" fill="#1baf7a"/>
<text x="104" y="82" font-size="13" fill="#0b0b0b">Sage-A1.5 Sangreal ICE 2.0</text>
<text x="830" y="82" font-size="12" fill="#898781" text-anchor="end">bars: ±1 s.e.</text>
</svg>

	<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 860 540" width="860" height="540" font-family="Inter,Segoe UI,Helvetica,Arial,sans-serif">
	<rect width="860" height="540" fill="#fcfcfb"/>
	<text x="84" y="30" font-size="19" font-weight="600" fill="#0b0b0b">Sage-A1.5-Sangreal-ICE: KL divergence vs size (code, 16K context)</text>
	<text x="84" y="52" font-size="13" fill="#52514e">Mean KLD vs Sage-A1.5 F16 on the same tokens, llama-perplexity, n_ctx 16384, 8 chunks. Lower-left is better.</text>
	<line x1="84" x2="830" y1="470.0" y2="470.0" stroke="#e8e7e3"/>
	<text x="76" y="474.0" font-size="12" fill="#898781" text-anchor="end">0.000</text>
	<line x1="84" x2="830" y1="395.2" y2="395.2" stroke="#e8e7e3"/>
	<text x="76" y="399.2" font-size="12" fill="#898781" text-anchor="end">0.025</text>
	<line x1="84" x2="830" y1="320.4" y2="320.4" stroke="#e8e7e3"/>
	<text x="76" y="324.4" font-size="12" fill="#898781" text-anchor="end">0.050</text>
	<line x1="84" x2="830" y1="245.6" y2="245.6" stroke="#e8e7e3"/>
	<text x="76" y="249.6" font-size="12" fill="#898781" text-anchor="end">0.075</text>
	<line x1="84" x2="830" y1="170.8" y2="170.8" stroke="#e8e7e3"/>
	<text x="76" y="174.8" font-size="12" fill="#898781" text-anchor="end">0.100</text>
	<line x1="84" x2="830" y1="96.0" y2="96.0" stroke="#e8e7e3"/>
	<text x="76" y="100.0" font-size="12" fill="#898781" text-anchor="end">0.125</text>
	<line x1="84.0" x2="84.0" y1="470" y2="475" stroke="#898781"/>
	<text x="84.0" y="490" font-size="12" fill="#898781" text-anchor="middle">10</text>
	<line x1="177.2" x2="177.2" y1="470" y2="475" stroke="#898781"/>
	<text x="177.2" y="490" font-size="12" fill="#898781" text-anchor="middle">12</text>
	<line x1="270.5" x2="270.5" y1="470" y2="475" stroke="#898781"/>
	<text x="270.5" y="490" font-size="12" fill="#898781" text-anchor="middle">14</text>
	<line x1="363.8" x2="363.8" y1="470" y2="475" stroke="#898781"/>
	<text x="363.8" y="490" font-size="12" fill="#898781" text-anchor="middle">16</text>
	<line x1="457.0" x2="457.0" y1="470" y2="475" stroke="#898781"/>
	<text x="457.0" y="490" font-size="12" fill="#898781" text-anchor="middle">18</text>
	<line x1="550.2" x2="550.2" y1="470" y2="475" stroke="#898781"/>
	<text x="550.2" y="490" font-size="12" fill="#898781" text-anchor="middle">20</text>
	<line x1="643.5" x2="643.5" y1="470" y2="475" stroke="#898781"/>
	<text x="643.5" y="490" font-size="12" fill="#898781" text-anchor="middle">22</text>
	<line x1="736.8" x2="736.8" y1="470" y2="475" stroke="#898781"/>
	<text x="736.8" y="490" font-size="12" fill="#898781" text-anchor="middle">24</text>
	<line x1="830.0" x2="830.0" y1="470" y2="475" stroke="#898781"/>
	<text x="830.0" y="490" font-size="12" fill="#898781" text-anchor="middle">26</text>
	<line x1="84" x2="830" y1="470" y2="470" stroke="#898781"/>
	<text x="457.0" y="518" font-size="13" fill="#52514e" text-anchor="middle">Model size without the MTP draft head (GB, lower is better)</text>
	<text transform="translate(24,283.0) rotate(-90)" font-size="13" fill="#52514e" text-anchor="middle">Mean KL divergence (lower is better)</text>
	<path d="M172.4,107.0 L294.5,249.1 L479.0,366.0 L667.6,387.6 L761.5,395.1" fill="none" stroke="#1baf7a" stroke-width="2"/>
	<line x1="172.4" x2="172.4" y1="115.2" y2="98.8" stroke="#1baf7a" stroke-width="1.5"/>
	<polygon points="172.4,100.0 178.9,112.0 165.9,112.0" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
	<text x="182.4" y="98.0" font-size="12" fill="#0b0b0b" text-anchor="start">12G-ICEv2 <tspan fill="#52514e">0.1213</tspan></text>
	<line x1="294.5" x2="294.5" y1="255.5" y2="242.6" stroke="#1baf7a" stroke-width="1.5"/>
	<polygon points="294.5,242.1 301.0,254.1 288.0,254.1" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
	<text x="304.5" y="240.1" font-size="12" fill="#0b0b0b" text-anchor="start">15G-ICEv2 <tspan fill="#52514e">0.0738</tspan></text>
	<line x1="479.0" x2="479.0" y1="370.5" y2="361.6" stroke="#1baf7a" stroke-width="1.5"/>
	<polygon points="479.0,359.0 485.5,371.0 472.5,371.0" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
	<text x="489.0" y="357.0" font-size="12" fill="#0b0b0b" text-anchor="start">19G-ICEv2 <tspan fill="#52514e">0.0347</tspan></text>
	<line x1="667.6" x2="667.6" y1="391.9" y2="383.3" stroke="#1baf7a" stroke-width="1.5"/>
	<polygon points="667.6,380.6 674.1,392.6 661.1,392.6" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
	<text x="677.6" y="378.6" font-size="12" fill="#0b0b0b" text-anchor="start">23G-ICEv2 <tspan fill="#52514e">0.0275</tspan></text>
	<line x1="761.5" x2="761.5" y1="399.3" y2="390.9" stroke="#1baf7a" stroke-width="1.5"/>
	<polygon points="761.5,388.1 768.0,400.1 755.0,400.1" fill="#1baf7a" stroke="#fcfcfb" stroke-width="1.5"/>
	<text x="761.5" y="417.1" font-size="12" fill="#0b0b0b" text-anchor="middle">25G-ICEv2 <tspan fill="#52514e">0.0250</tspan></text>
	<polygon points="90,72 96.5,84 83.5,84" fill="#1baf7a"/>
	<text x="104" y="82" font-size="13" fill="#0b0b0b">Sage-A1.5 Sangreal ICE 2.0</text>
	<text x="830" y="82" font-size="12" fill="#898781" text-anchor="end">bars: ±1 s.e.</text>
	</svg>