Instructions to use dispatchAI/SmolLM2-135M-Instruct-mobile with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dispatchAI/SmolLM2-135M-Instruct-mobile with Transformers:

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

pipe = pipeline("text-generation", model="dispatchAI/SmolLM2-135M-Instruct-mobile")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("dispatchAI/SmolLM2-135M-Instruct-mobile", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use dispatchAI/SmolLM2-135M-Instruct-mobile 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 dispatchAI/SmolLM2-135M-Instruct-mobile
# Run inference directly in the terminal:
llama cli -hf dispatchAI/SmolLM2-135M-Instruct-mobile

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf dispatchAI/SmolLM2-135M-Instruct-mobile
# Run inference directly in the terminal:
llama cli -hf dispatchAI/SmolLM2-135M-Instruct-mobile

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf dispatchAI/SmolLM2-135M-Instruct-mobile
# Run inference directly in the terminal:
./build/bin/llama cli -hf dispatchAI/SmolLM2-135M-Instruct-mobile

Use Docker

docker model run hf.co/dispatchAI/SmolLM2-135M-Instruct-mobile

Llama new
LM Studio
Jan

vLLM

How to use dispatchAI/SmolLM2-135M-Instruct-mobile with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "dispatchAI/SmolLM2-135M-Instruct-mobile"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "dispatchAI/SmolLM2-135M-Instruct-mobile",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/dispatchAI/SmolLM2-135M-Instruct-mobile

SGLang

How to use dispatchAI/SmolLM2-135M-Instruct-mobile 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 "dispatchAI/SmolLM2-135M-Instruct-mobile" \
    --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": "dispatchAI/SmolLM2-135M-Instruct-mobile",
		"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 "dispatchAI/SmolLM2-135M-Instruct-mobile" \
        --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": "dispatchAI/SmolLM2-135M-Instruct-mobile",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use dispatchAI/SmolLM2-135M-Instruct-mobile with Ollama:
```
ollama run hf.co/dispatchAI/SmolLM2-135M-Instruct-mobile
```
Unsloth Desktop
Docker Model Runner
How to use dispatchAI/SmolLM2-135M-Instruct-mobile with Docker Model Runner:
```
docker model run hf.co/dispatchAI/SmolLM2-135M-Instruct-mobile
```

Lemonade

How to use dispatchAI/SmolLM2-135M-Instruct-mobile with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull dispatchAI/SmolLM2-135M-Instruct-mobile

Run and chat with the model

lemonade run user.SmolLM2-135M-Instruct-mobile-{{QUANT_TAG}}

List all available models

lemonade list

Atomic Chat

3morixd commited on Jun 26

Commit

9443cc4

verified ·

1 Parent(s): 4f6669f

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +16 -22

README.md CHANGED Viewed

@@ -14,22 +14,19 @@ pipeline_tag: text-generation
 # SmolLM2-135M-Instruct-mobile
-A 135M parameter instruction-tuned model optimized for mobile and edge devices. 101MB GGUF file that runs at 59.7 tokens/sec on CPU.
-## ✅ Verified
-This model has been tested and verified working:
-- **Chat format**: `llama-3` (use `chat_format="llama-3"` in llama-cpp-python)
-- **Test date**: June 2026
-- **Verified with**: llama-cpp-python on CPU (112 cores, 128GB RAM)
-### Test Results
-| Prompt | Response | Correct? |
-|--------|----------|----------|
-| "What is the capital of France?" | "The capital of France is Paris." | ✅ |
-| "What is 2+2?" | "2 + 2 = 4" | ✅ |
-| "Write a one-sentence greeting." | "Welcome! I'm an AI assistant..." | ✅ |
 ## Model Details
@@ -40,15 +37,15 @@ This model has been tested and verified working:
 | **File Size** | 101 MB |
 | **Format** | GGUF |
 | **Chat Format** | llama-3 |
 | **License** | Apache-2.0 |
 ## Usage
 ### Python (llama-cpp-python)
 ```python
 from llama_cpp import Llama
 llm = Llama(model_path="model.gguf", chat_format="llama-3", n_ctx=512, n_threads=4)
 response = llm.create_chat_completion(
     messages=[{"role": "user", "content": "What is the capital of France?"}],
@@ -58,24 +55,21 @@ print(response["choices"][0]["message"]["content"])
 ```
 ### dispatchAI SDK
 ```python
 from dispatchai import load_model
 model = load_model("SmolLM2-135M-Instruct-mobile", backend="gguf")
 print(model.chat("What is the capital of France?"))
 ```
-### llama.cpp CLI
 ```bash
-llama-cli -m model.gguf -p "What is the capital of France?" -n 50 --chat-format llama-3
 ```
 ## Limitations
-- 135M parameters — best for simple tasks (QA, classification, short summaries)
-- Not suitable for complex reasoning or long-form generation
-- May repeat itself on some prompts
 - English-only
 ## About dispatchAI

 # SmolLM2-135M-Instruct-mobile
+A 135M parameter instruction-tuned model optimized for mobile. **101MB** GGUF file, **46.0 tokens/sec** on Snapdragon 865.
+## ✅ Verified on Real Phone Hardware
+Tested on Samsung S20 FE (Snapdragon 865, 8GB RAM, Android 13) using llama.cpp b9775:
+| Prompt | Response | Correct? | Speed |
+|--------|----------|----------|-------|
+| "The capital of France is" | "...Paris, which is a city in France" | ✅ | 46.0 t/s gen, 202.4 t/s prompt |
+| "What is 2+2?" | "2 + 2 = 4" | ✅ | — |
+| "Write a greeting" | "Welcome! I'm an AI assistant..." | ✅ | — |
+**Chat format**: `llama-3` — use `chat_format="llama-3"` in llama-cpp-python
 ## Model Details
 | **File Size** | 101 MB |
 | **Format** | GGUF |
 | **Chat Format** | llama-3 |
+| **Phone Speed** | 46.0 tokens/sec (Snapdragon 865) |
+| **CPU Speed** | 59.7 tokens/sec (112-core x86) |
 | **License** | Apache-2.0 |
 ## Usage
 ### Python (llama-cpp-python)
 ```python
 from llama_cpp import Llama
 llm = Llama(model_path="model.gguf", chat_format="llama-3", n_ctx=512, n_threads=4)
 response = llm.create_chat_completion(
     messages=[{"role": "user", "content": "What is the capital of France?"}],
 ```
 ### dispatchAI SDK
 ```python
 from dispatchai import load_model
 model = load_model("SmolLM2-135M-Instruct-mobile", backend="gguf")
 print(model.chat("What is the capital of France?"))
 ```
+### On Android (via ADB)
 ```bash
+adb push model.gguf /data/local/tmp/
+adb shell "cd /data/local/tmp && LD_LIBRARY_PATH=/data/local/tmp ./llama-cli -m model.gguf -p 'Hello' -n 30 -t 4 -st"
 ```
 ## Limitations
+- 135M params — best for simple tasks (QA, classification, short summaries)
+- May repeat on some prompts
 - English-only
 ## About dispatchAI