Instructions to use igorls/gemma4-e4b-classifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use igorls/gemma4-e4b-classifier with Transformers:

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

pipe = pipeline("text-generation", model="igorls/gemma4-e4b-classifier")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("igorls/gemma4-e4b-classifier")
model = AutoModelForCausalLM.from_pretrained("igorls/gemma4-e4b-classifier")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use igorls/gemma4-e4b-classifier with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="igorls/gemma4-e4b-classifier",
	filename="gemma4-e4b-classifier-Q4_K_M.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use igorls/gemma4-e4b-classifier with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf igorls/gemma4-e4b-classifier:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf igorls/gemma4-e4b-classifier:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf igorls/gemma4-e4b-classifier:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf igorls/gemma4-e4b-classifier: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 igorls/gemma4-e4b-classifier:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf igorls/gemma4-e4b-classifier: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 igorls/gemma4-e4b-classifier:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf igorls/gemma4-e4b-classifier:Q4_K_M

Use Docker

docker model run hf.co/igorls/gemma4-e4b-classifier:Q4_K_M

LM Studio
Jan

vLLM

How to use igorls/gemma4-e4b-classifier with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "igorls/gemma4-e4b-classifier"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "igorls/gemma4-e4b-classifier",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/igorls/gemma4-e4b-classifier:Q4_K_M

SGLang

How to use igorls/gemma4-e4b-classifier 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 "igorls/gemma4-e4b-classifier" \
    --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": "igorls/gemma4-e4b-classifier",
		"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 "igorls/gemma4-e4b-classifier" \
        --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": "igorls/gemma4-e4b-classifier",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use igorls/gemma4-e4b-classifier with Ollama:
```
ollama run hf.co/igorls/gemma4-e4b-classifier:Q4_K_M
```

Unsloth Studio new

How to use igorls/gemma4-e4b-classifier 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 igorls/gemma4-e4b-classifier 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 igorls/gemma4-e4b-classifier to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for igorls/gemma4-e4b-classifier to start chatting

Pi new

How to use igorls/gemma4-e4b-classifier with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf igorls/gemma4-e4b-classifier: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": "igorls/gemma4-e4b-classifier:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use igorls/gemma4-e4b-classifier with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf igorls/gemma4-e4b-classifier: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 igorls/gemma4-e4b-classifier:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use igorls/gemma4-e4b-classifier with Docker Model Runner:
```
docker model run hf.co/igorls/gemma4-e4b-classifier:Q4_K_M
```

Lemonade

How to use igorls/gemma4-e4b-classifier with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull igorls/gemma4-e4b-classifier:Q4_K_M

Run and chat with the model

lemonade run user.gemma4-e4b-classifier-Q4_K_M

List all available models

lemonade list

igorls commited on 3 days ago

Commit

61d9293

verified ·

1 Parent(s): a3624e6

docs: reframe model card around the model itself (drop project-specific shipping framing)

Browse files

Files changed (1) hide show

README.md +6 -21

README.md CHANGED Viewed

@@ -21,7 +21,7 @@ A modality-stripped variant of [`google/gemma-4-E4B-it`](https://huggingface.co/
 **Headline:** Same instruction-tuned text behavior as the official Gemma 4 E4B-it — including its multilingual coverage — but at **6.5 GB resident VRAM instead of 10.6 GB** (Ollama Q4_K_M, RTX 3090, Linux). All safety alignment is preserved — this is **not** an abliterated or uncensored variant.
-**For 8 GB GPU users:** this is the recommended Gemma 4 E4B variant. The official `gemma4:e4b-it-q4_K_M` does not fit on 8 GB cards even at short contexts (10.2 GB resident at ctx=8192). This variant fits with ~2 GB headroom and preserves every measured capability of the base model.
 ## Why this exists
@@ -47,32 +47,17 @@ All accuracy deltas are within statistical noise at n=100. The 4.1 GB VRAM win i
 ## Multilingual robustness
-The strip preserves Gemma 4's multilingual capability. Measured against the MemPalace harness translated to Portuguese (pt-BR), Spanish (es), and Chinese (zh), at parity context (ctx=8192) and with a multilingual scoring embedding (`embeddinggemma`) so cross-lingual cosine isn't penalized by the EN-only `nomic-embed-text` v1:
 | Task | en | pt-BR | es | zh |
 |---|---:|---:|---:|---:|
 | Calibration | 1.000 | 0.950 | 0.950 | 0.950 |
-| Room classification (closed) | 0.624 | 0.584 | 0.584 | 0.584 |
-| Room classification (open) ★ | **0.676** | **0.636** | **0.641** | **0.639** |
 | Entity extraction (F1) | 0.732 | 0.747 | 0.747 | 0.694 |
 | Memory coverage | 0.912 | 0.850 | 0.850 | 0.912 |
-This model is the **most language-stable** of the four 4B-class local candidates evaluated — closed/open room classification stays within ±0.02 across languages, where competing Qwen 3 variants degrade visibly on zh (closed-set drops to 0.535 for `qwen3:4b-instruct-2507-q8_0`).
-### When to pick this model vs Qwen 3 4B alternatives
-Same harness, same matrix, ctx=8192, full datasets:
-| Capability | Winner | Notes |
-|---|---|---|
-| **Open-set room classification** ★ | **this model** | 0.636-0.676 across 4 languages vs Qwen 0.56-0.63. The unique Gemma 4 strength replicating across every language tested. |
-| Closed-set room classification | rough tie | This model and `qwen3.5:4b-q4_K_M` trade the lead by 1-3 points. |
-| Memory extraction | rough tie (~0.85) | This model, `qwen3:4b-instruct-2507-q8_0`, and official Gemma 4 within 0.02 of each other. |
-| Entity extraction (F1) | Qwen 3 4B Q8 | `qwen3:4b-instruct-2507-q8_0` leads by 5-7 points on entity extraction across all 4 languages. |
-| TPS (output throughput) | Qwen 3 4B Q8 | 2x faster (220+ TPS vs ~130 TPS at ctx=8192). |
-| VRAM resident at ctx=8192 | rough tie | This model 6.1 GB, qwen3:4b-q8 5.8 GB, qwen3.5:4b-q4 6.0 GB. |
-**Pick this model** when slug-quality matters (open-set room routing / "what room does this conversation go in" UX) or when multilingual stability matters. **Pick `qwen3:4b-instruct-2507-q8_0`** when speed matters more than open-set slug quality, or when entity extraction is the dominant load.
 ## What was actually dropped
@@ -167,7 +152,7 @@ The official MTP drafter [`google/gemma-4-E4B-it-assistant`](https://huggingface
 | JSON entity list (128 tok) | 128 | 12291 ms | 6712 ms | 1.83x |
 | JSON memories (114 tok) | 114 | 8425 ms | **2771 ms** | **3.04x** |
-Speedup tracks output predictability — structured JSON outputs (the most common MemPalace surface) land at the high end (3x).
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer

 **Headline:** Same instruction-tuned text behavior as the official Gemma 4 E4B-it — including its multilingual coverage — but at **6.5 GB resident VRAM instead of 10.6 GB** (Ollama Q4_K_M, RTX 3090, Linux). All safety alignment is preserved — this is **not** an abliterated or uncensored variant.
+Fits comfortably on **8 GB GPUs at Q4_K_M** with realistic context lengths (5.85 GB resident at ctx=4096, 5.96 GB at ctx=8192). The official multimodal Q4_K_M sits at 10.2 GB resident even at ctx=8192 and won't load on 8 GB cards.
 ## Why this exists
 ## Multilingual robustness
+The strip preserves the base model's multilingual capability. Same classification + extraction tasks were run with inputs translated into Portuguese (pt-BR), Spanish (es), and Chinese (zh) — labels and the slug taxonomy kept in English to test the realistic cross-lingual mapping case. Scoring uses `embeddinggemma` for semantic similarity so cross-lingual cosine isn't artificially penalized.
 | Task | en | pt-BR | es | zh |
 |---|---:|---:|---:|---:|
 | Calibration | 1.000 | 0.950 | 0.950 | 0.950 |
+| Room classification (closed-set) | 0.624 | 0.584 | 0.584 | 0.584 |
+| Room classification (open-set) | 0.676 | 0.636 | 0.641 | 0.639 |
 | Entity extraction (F1) | 0.732 | 0.747 | 0.747 | 0.694 |
 | Memory coverage | 0.912 | 0.850 | 0.850 | 0.912 |
+Closed/open room classification stays within ±0.02 across all four languages; entity F1 within ±0.05; memory coverage within ±0.06. The strip did not introduce a multilingual regression. Models still emit responses in the input language by default — if your application needs same-language extraction (e.g. memories phrased in Portuguese for Portuguese conversations), the model does that natively.
 ## What was actually dropped
 | JSON entity list (128 tok) | 128 | 12291 ms | 6712 ms | 1.83x |
 | JSON memories (114 tok) | 114 | 8425 ms | **2771 ms** | **3.04x** |
+Speedup tracks output predictability — structured JSON outputs land at the high end (3x), short slug/letter classifications around 1.5-2x, free-form continuations near 1x.
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer