Instructions to use Akahsizrr/fuse-1-Lite-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M

Use Docker

docker model run hf.co/Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use Akahsizrr/fuse-1-Lite-GGUF with Ollama:
```
ollama run hf.co/Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
```
Unsloth Desktop

How to use Akahsizrr/fuse-1-Lite-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Akahsizrr/fuse-1-Lite-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": "Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Akahsizrr/fuse-1-Lite-GGUF with Docker Model Runner:
```
docker model run hf.co/Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M
```

Lemonade

How to use Akahsizrr/fuse-1-Lite-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.fuse-1-Lite-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-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 Akahsizrr/fuse-1-Lite-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Akahsizrr/fuse-1-Lite-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Akahsizrr/fuse-1-Lite-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 "Akahsizrr/fuse-1-Lite-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"

Integration issue

by DarkKitsune - opened Aug 7

Discussion

DarkKitsune

Aug 7

•

edited Aug 7

Hi, I tried following the instructions in INTEGRATION.md to integrate FUSE3 into a local fork of llama.cpp (based on b10235 specifically, a release from 5 days ago). I can make it as far as step 3, until I try to modify LLM_TENSOR_NAMES as instructed. However, the map in this version of llama.cpp is structured in a way that I cannot insert the provided code into it:

static const std::map<llm_tensor, const char *> LLM_TENSOR_NAMES = {
    { LLM_TENSOR_TOKEN_EMBD,                             "token_embd" },
    { LLM_TENSOR_OUTPUT_NORM,                            "output_norm" },
    { LLM_TENSOR_OUTPUT_NORM_LFM2,                       "token_embd_norm" }, // fix for wrong tensor name
    { LLM_TENSOR_OUTPUT,                                 "output" },
    { LLM_TENSOR_ROPE_FREQS,                             "rope_freqs" },
    { LLM_TENSOR_ATTN_NORM,                              "blk.%d.attn_norm" },
    { LLM_TENSOR_ATTN_Q,                                 "blk.%d.attn_q" },
...
    { LLM_TENSOR_DSPARK_CONF_PROJ,                       "conf_proj" },
};

But INTEGRATION.md says I need to insert this:

// In LLM_TENSOR_NAMES for FUSE3:
{
    LLM_ARCH_FUSE3,
    {
        { LLM_TENSOR_TOKEN_EMBD,      "token_embd.weight" },
        { LLM_TENSOR_OUTPUT_NORM_LFM2, "token_embd_norm.weight" },
        { LLM_TENSOR_OUTPUT,          "output.weight" },
        { LLM_TENSOR_ATTN_NORM,       "blk.{bid}.attn_norm.weight" },
        { LLM_TENSOR_ATTN_Q,          "blk.{bid}.attn_q.weight" },
        { LLM_TENSOR_ATTN_K,          "blk.{bid}.attn_k.weight" },
        { LLM_TENSOR_ATTN_V,          "blk.{bid}.attn_v.weight" },
        { LLM_TENSOR_ATTN_Q_NORM,     "blk.{bid}.attn_q_norm.weight" },
        { LLM_TENSOR_ATTN_K_NORM,     "blk.{bid}.attn_k_norm.weight" },
        { LLM_TENSOR_ATTN_OUT,        "blk.{bid}.attn_output.weight" },
        { LLM_TENSOR_FFN_NORM,        "blk.{bid}.ffn_norm.weight" },
        { LLM_TENSOR_FFN_GATE,        "blk.{bid}.ffn_gate.weight" },
        { LLM_TENSOR_FFN_DOWN,        "blk.{bid}.ffn_down.weight" },
        { LLM_TENSOR_FFN_UP,          "blk.{bid}.ffn_up.weight" },
        { LLM_TENSOR_SHORTCONV_CONV,    "blk.{bid}.shortconv_conv.weight" },
        { LLM_TENSOR_SHORTCONV_INPROJ,  "blk.{bid}.shortconv_inproj.weight" },
        { LLM_TENSOR_SHORTCONV_OUTPROJ, "blk.{bid}.shortconv_outproj.weight" },
    }
},

Do I need an older version of llama.cpp perhaps? Or am I simply misunderstanding?

Akahsizrr

Owner Aug 7

this is becasue it uses a custom architecure, llama integration insnt easy, am working on it today to make it way simpler

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment