Instructions to use 4-alokk/flashflow-gemma4-e2b-bank-sms with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use 4-alokk/flashflow-gemma4-e2b-bank-sms 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 4-alokk/flashflow-gemma4-e2b-bank-sms
# Run inference directly in the terminal:
llama cli -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf 4-alokk/flashflow-gemma4-e2b-bank-sms
# Run inference directly in the terminal:
llama cli -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

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 4-alokk/flashflow-gemma4-e2b-bank-sms
# Run inference directly in the terminal:
./llama-cli -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

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 4-alokk/flashflow-gemma4-e2b-bank-sms
# Run inference directly in the terminal:
./build/bin/llama-cli -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

Use Docker

docker model run hf.co/4-alokk/flashflow-gemma4-e2b-bank-sms

LM Studio
Jan
Ollama
How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with Ollama:
```
ollama run hf.co/4-alokk/flashflow-gemma4-e2b-bank-sms
```
Unsloth Desktop

How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

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": "4-alokk/flashflow-gemma4-e2b-bank-sms"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with Docker Model Runner:
```
docker model run hf.co/4-alokk/flashflow-gemma4-e2b-bank-sms
```

Lemonade

How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull 4-alokk/flashflow-gemma4-e2b-bank-sms

Run and chat with the model

lemonade run user.flashflow-gemma4-e2b-bank-sms-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

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 4-alokk/flashflow-gemma4-e2b-bank-sms

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use 4-alokk/flashflow-gemma4-e2b-bank-sms with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf 4-alokk/flashflow-gemma4-e2b-bank-sms

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 "4-alokk/flashflow-gemma4-e2b-bank-sms" \
  --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"

FlashFlow Gemma 4 E2B — Bank SMS Extraction

A LoRA fine-tuned Gemma 4 E2B (google/gemma-4-E2B-it) that reads raw Indian bank SMS and generates structured 13-field JSON transaction records, merged and quantized to GGUF for on-device inference via llama.cpp.

Built for the Build with Gemma: TFUG Prayagraj hackathon, powering the FlashFlow Sentinel Android app — a 100% offline financial SMS tracker.

Files

gemma4_e2b_q4km.gguf — Q4_K_M quantized, ~3.4GB. This is what the FlashFlow Sentinel app bundles (split into <1GB chunks at build time for Flutter asset loading).

Usage

{
  "transaction_type": "debit",
  "amount": 449.0,
  "currency": "INR",
  "date": "2026-06-14",
  "time": "19:32",
  "sender_bank": "HDFC Bank",
  "sender_acc": "XX1234",
  "receiver_bank": null,
  "receiver_acc": null,
  "counterparty_name": "SWIGGY",
  "reference_id": "615243987012",
  "balance_after": 18230.55,
  "is_actionable": true
}

Prompt format (Gemma's native turn format):

<start_of_turn>user
You are a financial data extractor. Parse the bank SMS and return a JSON object with these fields: transaction_type, amount, currency, date (ISO 8601), time, sender_bank, sender_acc, receiver_bank, receiver_acc, counterparty_name, reference_id, balance_after, is_actionable. Use null for absent fields.

SMS: <raw SMS text><end_of_turn>
<start_of_turn>model

Training

LoRA rank/alpha/dropout: 8/16/0.05, target modules scoped to model.language_model.layers.*.(self_attn|mlp).*_proj
100 steps, effective batch 8, bf16, on a 2,500-sample synthetic dataset
Full recipe, dataset generator, and training code: https://github.com/4-alok/flashflow-gemma4

Dataset

Synthetic Indian bank SMS, same generation methodology as 4-alokk/flashflow-bank-sms.

Downloads last month: 48

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for 4-alokk/flashflow-gemma4-e2b-bank-sms

Base model

google/gemma-4-E2B

Finetuned

google/gemma-4-E2B-it

Adapter

(187)

this model