Instructions to use ashaibani/slipstream-minicpm5-1b-forecaster with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ashaibani/slipstream-minicpm5-1b-forecaster with TimesFM:

# No code snippets available yet for this library.

# To use this model, check the repository files and the library's documentation.

# Want to help? PRs adding snippets are welcome at:
# https://github.com/huggingface/huggingface.js

llama-cpp-python

How to use ashaibani/slipstream-minicpm5-1b-forecaster with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ashaibani/slipstream-minicpm5-1b-forecaster",
	filename="minicpm5-1b-slipstream-q4_k_m.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use ashaibani/slipstream-minicpm5-1b-forecaster with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf ashaibani/slipstream-minicpm5-1b-forecaster: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 ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ashaibani/slipstream-minicpm5-1b-forecaster: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 ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Use Docker

docker model run hf.co/ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

LM Studio
Jan

vLLM

How to use ashaibani/slipstream-minicpm5-1b-forecaster with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ashaibani/slipstream-minicpm5-1b-forecaster"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ashaibani/slipstream-minicpm5-1b-forecaster",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Ollama
How to use ashaibani/slipstream-minicpm5-1b-forecaster with Ollama:
```
ollama run hf.co/ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
```

Unsloth Studio

How to use ashaibani/slipstream-minicpm5-1b-forecaster with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ashaibani/slipstream-minicpm5-1b-forecaster to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for ashaibani/slipstream-minicpm5-1b-forecaster to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ashaibani/slipstream-minicpm5-1b-forecaster to start chatting

How to use ashaibani/slipstream-minicpm5-1b-forecaster with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use ashaibani/slipstream-minicpm5-1b-forecaster with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf ashaibani/slipstream-minicpm5-1b-forecaster: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 ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use ashaibani/slipstream-minicpm5-1b-forecaster with Docker Model Runner:
```
docker model run hf.co/ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M
```

Lemonade

How to use ashaibani/slipstream-minicpm5-1b-forecaster with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ashaibani/slipstream-minicpm5-1b-forecaster:Q4_K_M

Run and chat with the model

lemonade run user.slipstream-minicpm5-1b-forecaster-Q4_K_M

List all available models

lemonade list

Slipstream · MiniCPM5-1B project-controls forecasting agent (GGUF)

A LoRA fine-tune of MiniCPM5-1B, distilled from Kimi K2.6's agentic traces, that acts as a project-controls forecasting agent. It runs a strict tool-calling loop - classical Earned-Value metrics + Google TimesFM 2.5 time-series forecasting - to project schedule slippage, final cost (EAC) and overrun risk, then reconciles the evidence into a final estimate with its own reasoning. Built for the Build Small Hackathon (Backyard AI track).

Results - held-out, vs the Kimi K2.6 parent, through the same agent loop

agent	valid_rate	finish-period err (median)	EAC error (median)
this model (MiniCPM5-1B distilled, LoRA r64)	1.00	3.5	1.8%
Kimi K2.6 (parent)	1.00	3.5	1.7%

The 1B student is statistically indistinguishable from its frontier parent on this task.

Run with llama.cpp

llama-cli -m minicpm5-1b-slipstream-q8_0.gguf -ngl 99 -c 8192 --jinja

Or via llama-cpp-python inside the agent loop - see the Slipstream repo (app + src/local_llm.py). Fully offline; no cloud APIs. Q8_0, ~1.15 GB.

Training

LoRA r64 on MiniCPM5-1B; 132 quality-filtered Kimi K2.6 traces + 35 curated project-controls domain Q&A, in the OpenAI tool-call message format (= the MiniCPM chat template directly). Distilled and exported on Modal; merge + GGUF conversion run on CPU.

Downloads last month: 323

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

8-bit

Model tree for ashaibani/slipstream-minicpm5-1b-forecaster

Base model

openbmb/MiniCPM5-1B

Quantized

(35)

this model

Quantizations

1 model

ashaibani
/

slipstream-minicpm5-1b-forecaster

Slipstream · MiniCPM5-1B project-controls forecasting agent (GGUF)

Results - held-out, vs the Kimi K2.6 parent, through the same agent loop

Run with llama.cpp

Training

Model tree for ashaibani/slipstream-minicpm5-1b-forecaster

Spaces using ashaibani/slipstream-minicpm5-1b-forecaster 2