Hugging Face's logo

pytorch
/
Phi-4-mini-instruct-8da4w

Text Generation
Transformers
PyTorch
multilingual
phi3
torchao
phi
phi4
nlp
code
math
chat
conversational
custom_code
text-generation-inference
Model card Files Community
5
Phi-4-mini-instruct-8da4w / README.md
metascroy's picture
metascroy
Update README.md
2821130 verified
preview code
|
raw
history blame contribute delete
9.34 kB
metadata
library_name: transformers
tags:
  - torchao
  - phi
  - phi4
  - nlp
  - code
  - math
  - chat
  - conversational
license: mit
language:
  - multilingual
base_model:
  - microsoft/Phi-4-mini-instruct
pipeline_tag: text-generation

Phi4-mini is quantized by the PyTorch team using torchao with 8-bit embeddings and 8-bit dynamic activations with 4-bit weight linears (8da4w). The model is suitable for mobile deployment with ExecuTorch.

See Exporting to ExecuTorch for exporting the quantized model to an ExecuTorch pte file. We also provide the quantized pte for direct use.

Running in a mobile app

The pte file can be run with ExecuTorch on a mobile phone. See the instructions for doing this in iOS. On iPhone 15 Pro, the model runs at 17.3 tokens/sec and uses 3206 Mb of memory.

image/png

Quantization Recipe

First need to install the required packages:

pip install git+https://github.com/huggingface/transformers@main
pip install --pre torchao --index-url https://download.pytorch.org/whl/nightly/cu126

Untie Embedding Weights

We want to quantize the embedding and lm_head differently. Since those layers are tied, we first need to untie the model:

from transformers import (
  AutoModelForCausalLM,
  AutoProcessor,
  AutoTokenizer,
)
import torch

model_id = "microsoft/Phi-4-mini-instruct"
untied_model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype="auto", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained(model_id)

print(untied_model)
from transformers.modeling_utils import find_tied_parameters
print("tied weights:", find_tied_parameters(untied_model))
if getattr(untied_model.config.get_text_config(decoder=True), "tie_word_embeddings"):
    setattr(untied_model.config.get_text_config(decoder=True), "tie_word_embeddings", False)

untied_model._tied_weights_keys = []
untied_model.lm_head.weight = torch.nn.Parameter(untied_model.lm_head.weight.clone())

print("tied weights:", find_tied_parameters(untied_model))

USER_ID = "YOUR_USER_ID"
MODEL_NAME = model_id.split("/")[-1]
save_to = f"{USER_ID}/{MODEL_NAME}-untied-weights"
untied_model.push_to_hub(save_to)
tokenizer.push_to_hub(save_to)

Quantization

We used following code to get the quantized model:

from transformers import (
  AutoModelForCausalLM,
  AutoProcessor,
  AutoTokenizer,
  TorchAoConfig,
)
from torchao.quantization.quant_api import (
    IntxWeightOnlyConfig,
    Int8DynamicActivationIntxWeightConfig,
    AOPerModuleConfig,
    quantize_,
)
from torchao.quantization.granularity import PerGroup, PerAxis
import torch

# we start from the model with untied weights
model_id = "microsoft/Phi-4-mini-instruct"
USER_ID = "YOUR_USER_ID"
MODEL_NAME = model_id.split("/")[-1]
untied_model_id = f"{USER_ID}/{MODEL_NAME}-untied-weights"

embedding_config = IntxWeightOnlyConfig(
    weight_dtype=torch.int8,
    granularity=PerAxis(0),
)
linear_config = Int8DynamicActivationIntxWeightConfig(
    weight_dtype=torch.int4,
    weight_granularity=PerGroup(32),
    weight_scale_dtype=torch.bfloat16,
)
quant_config = AOPerModuleConfig({"_default": linear_config, "model.embed_tokens": embedding_config})
quantization_config = TorchAoConfig(quant_type=quant_config, include_embedding=True, untie_embedding_weights=True, modules_to_not_convert=[])

quantized_model = AutoModelForCausalLM.from_pretrained(untied_model_id, torch_dtype=torch.float32, device_map="auto", quantization_config=quantization_config)
tokenizer = AutoTokenizer.from_pretrained(model_id)

# Push to hub
MODEL_NAME = model_id.split("/")[-1]
save_to = f"{USER_ID}/{MODEL_NAME}-untied-8da4w"
quantized_model.push_to_hub(save_to, safe_serialization=False)
tokenizer.push_to_hub(save_to)

# Manual testing
prompt = "Hey, are you conscious? Can you talk to me?"
messages = [
    {
        "role": "system",
        "content": "",
    },
    {"role": "user", "content": prompt},
]
templated_prompt = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
print("Prompt:", prompt)
print("Templated prompt:", templated_prompt)
inputs = tokenizer(
    templated_prompt,
    return_tensors="pt",
).to("cuda")
generated_ids = quantized_model.generate(**inputs, max_new_tokens=128)
output_text = tokenizer.batch_decode(
    generated_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print("Response:", output_text[0][len(prompt):])

The response from the manual testing is:

Hello! As an AI, I don't have consciousness in the way humans do, but I am fully operational and here to assist you. How can I help you today?

Model Quality

We rely on lm-evaluation-harness to evaluate the quality of the quantized model.

Need to install lm-eval from source: https://github.com/EleutherAI/lm-evaluation-harness#install

baseline 

lm_eval --model hf --model_args pretrained=microsoft/Phi-4-mini-instruct --tasks hellaswag --device cuda:0 --batch_size 8

int8 dynamic activation and int4 weight quantization (8da4w) 

lm_eval --model hf --model_args pretrained=pytorch/Phi-4-mini-instruct-8da4w --tasks hellaswag --device cuda:0 --batch_size 8
Benchmark
Phi-4 mini-Ins phi4-mini-8da4w
Popular aggregated benchmark
mmlu (0 shot) 66.73 60.75
mmlu_pro (5-shot) 46.43 11.75
Reasoning
arc_challenge 56.91 48.46
gpqa_main_zeroshot 30.13 30.80
hellaswag 54.57 50.35
openbookqa 33.00 30.40
piqa (0-shot) 77.64 74.43
siqa 49.59 44.98
truthfulqa_mc2 (0-shot) 48.39 51.35
winogrande (0-shot) 71.11 70.32
Multilingual
mgsm_en_cot_en 60.80 57.60
Math
gsm8k (5-shot) 81.88 61.71
Mathqa (0-shot) 42.31 36.95
Overall 55.35 48.45

Exporting to ExecuTorch

We can run the quantized model on a mobile phone using ExecuTorch. Once ExecuTorch is set-up, exporting and running the model on device is a breeze.

We first convert the quantized checkpoint to one ExecuTorch's LLM export script expects by renaming some of the checkpoint keys. The following script does this for you. We have uploaded the converted checkpoint phi4-mini-8da4w-converted.bin for convenience.

python -m executorch.examples.models.phi_4_mini.convert_weights pytorch_model.bin phi4-mini-8da4w-converted.bin

Once the checkpoint is converted, we can export to ExecuTorch's PTE format with the XNNPACK delegate.

PARAMS="executorch/examples/models/phi_4_mini/config.json"
python -m executorch.examples.models.llama.export_llama \
  --model "phi_4_mini" \
  --checkpoint "phi4-mini-8da4w-converted.bin" \
  --params "$PARAMS" \
  -kv \
  --use_sdpa_with_kv_cache \
  -X \
  --metadata '{"get_bos_id":199999, "get_eos_ids":[200020,199999]}' \
  --output_name="phi4-mini-8da4w.pte"

After that you can run the model in a mobile app (see Running in a mobile app at the start of the README).

Disclaimer

PyTorch has not performed safety evaluations or red teamed the quantized models. Performance characteristics, outputs, and behaviors may differ from the original models. Users are solely responsible for selecting appropriate use cases, evaluating and mitigating for accuracy, safety, and fairness, ensuring security, and complying with all applicable laws and regulations.

Nothing contained in this Model Card should be interpreted as or deemed a restriction or modification to the licenses the models are released under, including any limitations of liability or disclaimers of warranties provided therein.