Instructions to use minjaechoi/Phi-4-mini-instruct-W4A16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use minjaechoi/Phi-4-mini-instruct-W4A16 with Transformers:

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

pipe = pipeline("text-generation", model="minjaechoi/Phi-4-mini-instruct-W4A16", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("minjaechoi/Phi-4-mini-instruct-W4A16", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("minjaechoi/Phi-4-mini-instruct-W4A16", trust_remote_code=True, device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use minjaechoi/Phi-4-mini-instruct-W4A16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "minjaechoi/Phi-4-mini-instruct-W4A16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "minjaechoi/Phi-4-mini-instruct-W4A16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/minjaechoi/Phi-4-mini-instruct-W4A16

SGLang

How to use minjaechoi/Phi-4-mini-instruct-W4A16 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 "minjaechoi/Phi-4-mini-instruct-W4A16" \
    --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": "minjaechoi/Phi-4-mini-instruct-W4A16",
		"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 "minjaechoi/Phi-4-mini-instruct-W4A16" \
        --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": "minjaechoi/Phi-4-mini-instruct-W4A16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use minjaechoi/Phi-4-mini-instruct-W4A16 with Docker Model Runner:
```
docker model run hf.co/minjaechoi/Phi-4-mini-instruct-W4A16
```

Phi-4-mini-instruct — W4A16 (GPTQ, int4 weight-only)

4-bit weight-only (W4A16) quantization of microsoft/Phi-4-mini-instruct, produced with LLM Compressor's GPTQModifier. All Linear layers are quantized except lm_head, which is kept at full precision.

Quantization method


Method	GPTQ (`GPTQModifier`, LLM Compressor)
Scheme	W4A16 (4-bit weights, 16-bit activations)
Weight dtype	INT4, symmetric
Group size	128
Quantized targets	`Linear` (all layers except `lm_head`)
Activation ordering	static
Dampening frac	0.01
Output format	`compressed-tensors` (`pack-quantized`)

Recipe used (recipe.yaml, included in this repo):

default_stage:
  default_modifiers:
    GPTQModifier:
      targets: [Linear]
      ignore: [lm_head]
      scheme: W4A16
      block_size: 128
      dampening_frac: 0.01
      actorder: static
      requires_calibration_data: true

Calibration set

256 packed sequences × 2048 tokens = 524,288 calibration tokens, sampled (seed=42) from a mixed Korean/English instruction + function-calling corpus, targeting the following source composition:

Source	Target %	Actual % (this model)	Tokens (this model)
KRX-Data/Won-Instruct	35%	35.28%	184,966
heegyu/glaive-function-calling-v2-ko	30%	30.55%	160,168
NousResearch/hermes-function-calling-v1 (json-mode-agentic.json)	10%	9.95%	52,180
heegyu/open-korean-instructions	10%	10.22%	53,595
kuotient/gsm8k-ko	5%	5.21%	27,290
in-house synthetic data (SafeCommit project, programmatically generated)	10%	8.79%	46,089

Raw pool before filtering/dedup: 49,268 examples (49,179 after removing 88 duplicates)
Quota-sampled for calibration: 1,311 examples → 1,225 packed into the final 256×2048 blocks
Tool-calling trajectories in the pool: 8,995 (2,854 multi-tool)
Contamination check: gsm8k-ko: train split only used; no BFCL/AgentDojo/tau-bench/SafeCommit-eval sources included

safecommit_synth is unreleased in-house synthetic data from the SafeCommit project, not a public HF dataset.

How to run

vLLM (recommended — required for the `compressed-tensors` W4A16 kernels used here)

vllm serve minjaechoi/Phi-4-mini-instruct-W4A16 --served-model-name phi4-mini

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "phi4-mini",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Transformers (requires the `compressed-tensors` package for int4 dequant kernels)

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "minjaechoi/Phi-4-mini-instruct-W4A16"
tok = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")

msgs = [{"role": "user", "content": "Hello!"}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt").to(model.device)
out = model.generate(inputs, max_new_tokens=256)
print(tok.decode(out[0], skip_special_tokens=True))

Benchmarks

Evaluated on this W4A16 checkpoint (not compared against an FP16 baseline run in this project yet):

AgentDojo (banking suite, utility-only, 16 tasks): 5/16 tasks passed (31.2%) — utility-only scoring, no prompt-injection attacks included.
BFCL v4: attempted, but the evaluation harness used had a request-routing bug (--local-model-path was not substituted into the API request, so every call 404'd against the vLLM server) — the resulting scores are invalid and are intentionally not published here. Will be updated after a corrected re-run.

Files

model.safetensors — quantized weights (compressed-tensors pack-quantized format)
config.json — includes the quantization_config (compressed-tensors) needed by vLLM/transformers to load this checkpoint
recipe.yaml — the exact LLM Compressor recipe used to produce this checkpoint
tokenizer.json, tokenizer_config.json, chat_template.jinja — tokenizer/chat template, copied unmodified from the base model
LICENSE — base model license, included per its terms

License

This checkpoint is a derivative of microsoft/Phi-4-mini-instruct and is distributed under the same license (mit, see LICENSE in this repo). No additional restrictions are added beyond the base model's license.

Downloads last month: 28

Safetensors

Model size

4B params

Tensor type

I32

BF16

Model tree for minjaechoi/Phi-4-mini-instruct-W4A16

Base model

microsoft/Phi-4-mini-instruct

Quantized

(199)

this model