Instructions to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Transformers:

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

pipe = pipeline("text-generation", model="DuoNeural/Lattice-6-Prometheus-v2-HeptaCap")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import CDMMoELM
model = CDMMoELM.from_pretrained("DuoNeural/Lattice-6-Prometheus-v2-HeptaCap", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap 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 DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32
# Run inference directly in the terminal:
llama cli -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32
# Run inference directly in the terminal:
llama cli -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

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 DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32
# Run inference directly in the terminal:
./build/bin/llama cli -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

Use Docker

docker model run hf.co/DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

Llama new
LM Studio
Jan

vLLM

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

SGLang

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap 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 "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap" \
    --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": "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap",
		"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 "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap" \
        --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": "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Ollama:
```
ollama run hf.co/DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32
```
Unsloth Desktop

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

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": "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Docker Model Runner:
```
docker model run hf.co/DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32
```

Lemonade

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

Run and chat with the model

lemonade run user.Lattice-6-Prometheus-v2-HeptaCap-F32

List all available models

lemonade list

Hermes Agent

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

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 DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use DuoNeural/Lattice-6-Prometheus-v2-HeptaCap with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32

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 "DuoNeural/Lattice-6-Prometheus-v2-HeptaCap:F32" \
  --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"

DuoNeural commited on 3 days ago

Commit

04702fb

verified ·

1 Parent(s): 3f8760c

Add hybrid quantization metadata

Browse files

Files changed (1) hide show

lattice6_prometheus-q8-attn-q6-rest-text.gguf.quantization.json +182 -0

lattice6_prometheus-q8-attn-q6-rest-text.gguf.quantization.json ADDED Viewed

	@@ -0,0 +1,182 @@

+{
+  "input": "models/lattice6_prometheus-f32-text.gguf",
+  "output": "models/lattice6_prometheus-q8-attn-q6-rest-text.gguf",
+  "type": "q6_k",
+  "policy": "Q8_0 attention qkv/output; Q6_K other large matrices; F32 positions, norms, routing geometry, telepathy and mapping",
+  "input_bytes": 1634509344,
+  "output_bytes": 349445824,
+  "tensor_types": {
+    "F32": 68,
+    "Q6_K": 85,
+    "Q8_0": 14
+  },
+  "tensor_type_manifest": {
+    "token_embd.weight": "Q6_K",
+    "position_embd.weight": "F32",
+    "token_embd_norm.weight": "F32",
+    "token_embd_norm.bias": "F32",
+    "output_norm.weight": "F32",
+    "output_norm.bias": "F32",
+    "blk.0.attn_norm.weight": "F32",
+    "blk.0.attn_norm.bias": "F32",
+    "blk.0.ffn_norm.weight": "F32",
+    "blk.0.ffn_norm.bias": "F32",
+    "blk.0.attn_output.weight": "Q8_0",
+    "blk.0.attn_output.bias": "F32",
+    "blk.0.attn_qkv.weight": "Q8_0",
+    "blk.0.attn_qkv.bias": "F32",
+    "blk.0.cdm_slots.weight": "F32",
+    "blk.0.cdm_tele.weight": "F32",
+    "blk.0.cdm_slot_expert.weight": "F32",
+    "blk.0.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.0.cdm_expert_0_up.weight": "Q6_K",
+    "blk.0.cdm_expert_0_down.weight": "Q6_K",
+    "blk.0.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.0.cdm_expert_1_up.weight": "Q6_K",
+    "blk.0.cdm_expert_1_down.weight": "Q6_K",
+    "blk.0.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.0.cdm_expert_2_up.weight": "Q6_K",
+    "blk.0.cdm_expert_2_down.weight": "Q6_K",
+    "blk.0.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.0.cdm_expert_3_up.weight": "Q6_K",
+    "blk.0.cdm_expert_3_down.weight": "Q6_K",
+    "blk.1.attn_norm.weight": "F32",
+    "blk.1.attn_norm.bias": "F32",
+    "blk.1.ffn_norm.weight": "F32",
+    "blk.1.ffn_norm.bias": "F32",
+    "blk.1.attn_output.weight": "Q8_0",
+    "blk.1.attn_output.bias": "F32",
+    "blk.1.attn_qkv.weight": "Q8_0",
+    "blk.1.attn_qkv.bias": "F32",
+    "blk.1.cdm_slots.weight": "F32",
+    "blk.1.cdm_tele.weight": "F32",
+    "blk.1.cdm_slot_expert.weight": "F32",
+    "blk.1.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.1.cdm_expert_0_up.weight": "Q6_K",
+    "blk.1.cdm_expert_0_down.weight": "Q6_K",
+    "blk.1.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.1.cdm_expert_1_up.weight": "Q6_K",
+    "blk.1.cdm_expert_1_down.weight": "Q6_K",
+    "blk.1.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.1.cdm_expert_2_up.weight": "Q6_K",
+    "blk.1.cdm_expert_2_down.weight": "Q6_K",
+    "blk.1.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.1.cdm_expert_3_up.weight": "Q6_K",
+    "blk.1.cdm_expert_3_down.weight": "Q6_K",
+    "blk.2.attn_norm.weight": "F32",
+    "blk.2.attn_norm.bias": "F32",
+    "blk.2.ffn_norm.weight": "F32",
+    "blk.2.ffn_norm.bias": "F32",
+    "blk.2.attn_output.weight": "Q8_0",
+    "blk.2.attn_output.bias": "F32",
+    "blk.2.attn_qkv.weight": "Q8_0",
+    "blk.2.attn_qkv.bias": "F32",
+    "blk.2.cdm_slots.weight": "F32",
+    "blk.2.cdm_tele.weight": "F32",
+    "blk.2.cdm_slot_expert.weight": "F32",
+    "blk.2.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.2.cdm_expert_0_up.weight": "Q6_K",
+    "blk.2.cdm_expert_0_down.weight": "Q6_K",
+    "blk.2.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.2.cdm_expert_1_up.weight": "Q6_K",
+    "blk.2.cdm_expert_1_down.weight": "Q6_K",
+    "blk.2.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.2.cdm_expert_2_up.weight": "Q6_K",
+    "blk.2.cdm_expert_2_down.weight": "Q6_K",
+    "blk.2.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.2.cdm_expert_3_up.weight": "Q6_K",
+    "blk.2.cdm_expert_3_down.weight": "Q6_K",
+    "blk.3.attn_norm.weight": "F32",
+    "blk.3.attn_norm.bias": "F32",
+    "blk.3.ffn_norm.weight": "F32",
+    "blk.3.ffn_norm.bias": "F32",
+    "blk.3.attn_output.weight": "Q8_0",
+    "blk.3.attn_output.bias": "F32",
+    "blk.3.attn_qkv.weight": "Q8_0",
+    "blk.3.attn_qkv.bias": "F32",
+    "blk.3.cdm_slots.weight": "F32",
+    "blk.3.cdm_tele.weight": "F32",
+    "blk.3.cdm_slot_expert.weight": "F32",
+    "blk.3.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.3.cdm_expert_0_up.weight": "Q6_K",
+    "blk.3.cdm_expert_0_down.weight": "Q6_K",
+    "blk.3.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.3.cdm_expert_1_up.weight": "Q6_K",
+    "blk.3.cdm_expert_1_down.weight": "Q6_K",
+    "blk.3.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.3.cdm_expert_2_up.weight": "Q6_K",
+    "blk.3.cdm_expert_2_down.weight": "Q6_K",
+    "blk.3.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.3.cdm_expert_3_up.weight": "Q6_K",
+    "blk.3.cdm_expert_3_down.weight": "Q6_K",
+    "blk.4.attn_norm.weight": "F32",
+    "blk.4.attn_norm.bias": "F32",
+    "blk.4.ffn_norm.weight": "F32",
+    "blk.4.ffn_norm.bias": "F32",
+    "blk.4.attn_output.weight": "Q8_0",
+    "blk.4.attn_output.bias": "F32",
+    "blk.4.attn_qkv.weight": "Q8_0",
+    "blk.4.attn_qkv.bias": "F32",
+    "blk.4.cdm_slots.weight": "F32",
+    "blk.4.cdm_tele.weight": "F32",
+    "blk.4.cdm_slot_expert.weight": "F32",
+    "blk.4.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.4.cdm_expert_0_up.weight": "Q6_K",
+    "blk.4.cdm_expert_0_down.weight": "Q6_K",
+    "blk.4.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.4.cdm_expert_1_up.weight": "Q6_K",
+    "blk.4.cdm_expert_1_down.weight": "Q6_K",
+    "blk.4.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.4.cdm_expert_2_up.weight": "Q6_K",
+    "blk.4.cdm_expert_2_down.weight": "Q6_K",
+    "blk.4.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.4.cdm_expert_3_up.weight": "Q6_K",
+    "blk.4.cdm_expert_3_down.weight": "Q6_K",
+    "blk.5.attn_norm.weight": "F32",
+    "blk.5.attn_norm.bias": "F32",
+    "blk.5.ffn_norm.weight": "F32",
+    "blk.5.ffn_norm.bias": "F32",
+    "blk.5.attn_output.weight": "Q8_0",
+    "blk.5.attn_output.bias": "F32",
+    "blk.5.attn_qkv.weight": "Q8_0",
+    "blk.5.attn_qkv.bias": "F32",
+    "blk.5.cdm_slots.weight": "F32",
+    "blk.5.cdm_tele.weight": "F32",
+    "blk.5.cdm_slot_expert.weight": "F32",
+    "blk.5.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.5.cdm_expert_0_up.weight": "Q6_K",
+    "blk.5.cdm_expert_0_down.weight": "Q6_K",
+    "blk.5.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.5.cdm_expert_1_up.weight": "Q6_K",
+    "blk.5.cdm_expert_1_down.weight": "Q6_K",
+    "blk.5.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.5.cdm_expert_2_up.weight": "Q6_K",
+    "blk.5.cdm_expert_2_down.weight": "Q6_K",
+    "blk.5.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.5.cdm_expert_3_up.weight": "Q6_K",
+    "blk.5.cdm_expert_3_down.weight": "Q6_K",
+    "blk.6.attn_norm.weight": "F32",
+    "blk.6.attn_norm.bias": "F32",
+    "blk.6.ffn_norm.weight": "F32",
+    "blk.6.ffn_norm.bias": "F32",
+    "blk.6.attn_output.weight": "Q8_0",
+    "blk.6.attn_output.bias": "F32",
+    "blk.6.attn_qkv.weight": "Q8_0",
+    "blk.6.attn_qkv.bias": "F32",
+    "blk.6.cdm_slots.weight": "F32",
+    "blk.6.cdm_tele.weight": "F32",
+    "blk.6.cdm_slot_expert.weight": "F32",
+    "blk.6.cdm_expert_0_gate.weight": "Q6_K",
+    "blk.6.cdm_expert_0_up.weight": "Q6_K",
+    "blk.6.cdm_expert_0_down.weight": "Q6_K",
+    "blk.6.cdm_expert_1_gate.weight": "Q6_K",
+    "blk.6.cdm_expert_1_up.weight": "Q6_K",
+    "blk.6.cdm_expert_1_down.weight": "Q6_K",
+    "blk.6.cdm_expert_2_gate.weight": "Q6_K",
+    "blk.6.cdm_expert_2_up.weight": "Q6_K",
+    "blk.6.cdm_expert_2_down.weight": "Q6_K",
+    "blk.6.cdm_expert_3_gate.weight": "Q6_K",
+    "blk.6.cdm_expert_3_up.weight": "Q6_K",
+    "blk.6.cdm_expert_3_down.weight": "Q6_K"
+  }
+}