Instructions to use Harsha001/Sage-A1.5-ROCmFPX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Harsha001/Sage-A1.5-ROCmFPX 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 Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
llama cli -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
llama cli -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

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 Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
./llama-cli -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

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 Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Use Docker

docker model run hf.co/Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

LM Studio
Jan

vLLM

How to use Harsha001/Sage-A1.5-ROCmFPX with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Harsha001/Sage-A1.5-ROCmFPX"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Harsha001/Sage-A1.5-ROCmFPX",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Ollama

How to use Harsha001/Sage-A1.5-ROCmFPX with Ollama:

ollama run hf.co/Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Unsloth Desktop

How to use Harsha001/Sage-A1.5-ROCmFPX with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

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": "Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Harsha001/Sage-A1.5-ROCmFPX with Docker Model Runner:
```
docker model run hf.co/Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX
```

Lemonade

How to use Harsha001/Sage-A1.5-ROCmFPX with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Run and chat with the model

lemonade run user.Sage-A1.5-ROCmFPX-Q2_0_ROCMFPX

List all available models

lemonade list

Hermes Agent

How to use Harsha001/Sage-A1.5-ROCmFPX with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

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 Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Harsha001/Sage-A1.5-ROCmFPX with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX

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 "Harsha001/Sage-A1.5-ROCmFPX:Q2_0_ROCMFPX" \
  --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 — ROCmFPX GGUF

InternScience/Sage-A1.5, the 35B mixture-of-experts agentic model from Shanghai AI Laboratory, quantized to the ROCmFPX formats, a llama.cpp fork with low-bit weight types for AMD GPUs (Vulkan and HIP/ROCm).

Start with Q4_0_ROCMFP4_STRIX. It is the closest 4-bit file to the original.

These files only load in a ROCmFPX build of llama.cpp. Stock llama.cpp, LM Studio and Ollama do not know these types. These files have not been tested on AMD hardware yet.

Not an official release of InternScience. All credit for the model goes to its authors.

Files and measured quality

Measured against the BF16 model (72 GB, perplexity 6.2751) on WikiText-2 test, 40 chunks of 512 tokens. KLD is the mean KL divergence of the next-token distribution from BF16 (lower is better). Same top is how often the quant picks the same next token as BF16.

File	Size	PPL	KLD	Same top
`Sage-A1.5-Q8_0_ROCMFPX.gguf`	36.7 GB	6.2945	0.0107	95.62%	8-bit, closest to BF16
`Sage-A1.5-Q6_0_ROCMFPX.gguf`	29.4 GB	6.2826	0.0115	95.05%	6-bit, almost as close as 8-bit
`Sage-A1.5-Q4_0_ROCMFP4_STRIX.gguf`	19.1 GB	6.4536	0.0537	88.81%	Recommended. 4-bit recipe tuned for Strix Halo (`gfx1151`)
`Sage-A1.5-Q4_0_ROCMFP4_FAST.gguf`	19.0 GB	6.4873	0.0543	89.19%	speed-first 4-bit (one scale per block)
`Sage-A1.5-Q2_0_ROCMFPX.gguf`	11.4 GB	10.4152	0.5274	70.22%	smallest file; clear quality loss
`mmproj-Sage-A1.5-F16.gguf`	0.9 GB				vision projector (F16), converted but not tested

Not published: Q3_0_ROCMFPX (19.8 GB, PPL 6.6816, KLD 0.0741, same top 87.80%) is larger than Q4_0_ROCMFP4_STRIX and further from BF16.

The 8-bit file is not as close to BF16 as a stock Q8_0 usually is: ROCmFPX block scales have a fixed smallest step, so 8-bit blocks of small weights keep fewer than 8 bits (ROCmFPX#118).

How these were made

convert_hf_to_gguf.py --outtype bf16 (and --mmproj --outtype f16 for the vision projector).
Importance matrix from the BF16 model on the GPU over a general calibration text (calibration_datav3), 512-token chunks.
llama-quantize --imatrix, one run per type, from HarshaVardhan3002/ROCmFPX (ROCmFPX fb08d7c merged onto upstream llama.cpp).
Quality: the BF16 reference ran on an NVIDIA GPU with stock BF16 kernels. Each quant was expanded back to F16 (llama-quantize --allow-requantize ... F16), which keeps its stored weights, and scored on the same GPU with stock F16 kernels. The fork's own CUDA kernels for these types are not correct on NVIDIA, so they were not used. As a check, this method gave KLD 0.0050 for a 9B model's Q8_0_ROCMFPX where the CPU backend gave 0.0054. AMD GPU kernels compute the same math in a different order, so results on an AMD GPU can differ slightly.

Run

git clone https://github.com/HarshaVardhan3002/ROCmFPX && cd ROCmFPX
cmake -B build -DGGML_VULKAN=ON          # or -DGGML_HIP=ON -DAMDGPU_TARGETS=gfx1151 (gfx1100 for RX 7900)
cmake --build build --target llama-server -j
build/bin/llama-server -m Sage-A1.5-Q4_0_ROCMFP4_STRIX.gguf -ngl 999 -fa on --jinja

License

Apache-2.0, same as the base model.

Downloads last month: -

GGUF

Model size

36B params

Architecture

qwen35moe

Hardware compatibility

2-bit

4-bit

8-bit

View +1 variant

Model tree for Harsha001/Sage-A1.5-ROCmFPX

Base model

InternScience/Sage-A1.5

Quantized

(25)

this model