Instructions to use pszemraj/Phi-3-small-8k-prune6 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pszemraj/Phi-3-small-8k-prune6 with Transformers:

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

pipe = pipeline("text-generation", model="pszemraj/Phi-3-small-8k-prune6", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("pszemraj/Phi-3-small-8k-prune6", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use pszemraj/Phi-3-small-8k-prune6 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pszemraj/Phi-3-small-8k-prune6"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pszemraj/Phi-3-small-8k-prune6",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/pszemraj/Phi-3-small-8k-prune6

SGLang

How to use pszemraj/Phi-3-small-8k-prune6 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 "pszemraj/Phi-3-small-8k-prune6" \
    --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": "pszemraj/Phi-3-small-8k-prune6",
		"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 "pszemraj/Phi-3-small-8k-prune6" \
        --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": "pszemraj/Phi-3-small-8k-prune6",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use pszemraj/Phi-3-small-8k-prune6 with Docker Model Runner:
```
docker model run hf.co/pszemraj/Phi-3-small-8k-prune6
```

Phi-3-small-8k-instruct: 6 layers pruned

This is a layer-pruned language model created using mergekit. Layers to prune were selected based off of the average distances as follows:

Quick eval

Quick eval for: pszemraj/Phi-3-small-8k-prune6

hf (pretrained=pszemraj/Phi-3-small-8k-prune6,trust_remote_code=True,dtype=bfloat16), gen_kwargs: (None), limit: None, num_fewshot: None, batch_size: 2

Tasks	Version	Filter	Metric	Value		Stderr
arc_easy	1	none	acc	0.7479	±	0.0089
		none	acc_norm	0.7125	±	0.0093
boolq	2	none	acc	0.7489	±	0.0076
lambada_openai	1	none	perplexity	27.3270	±	1.0861
		none	acc	0.3600	±	0.0067
openbookqa	1	none	acc	0.3360	±	0.0211
		none	acc_norm	0.4020	±	0.0219
piqa	1	none	acc	0.7182	±	0.0105
		none	acc_norm	0.7329	±	0.0103
winogrande	1	none	acc	0.7143	±	0.0127

Usage

While some further pre-training will be good, it seems capable of generating coherent text as is.

from transformers import AutoModelForCausalLM, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained(
    "microsoft/Phi-3-small-8k-instruct", trust_remote_code=True
)
model = AutoModelForCausalLM.from_pretrained(
    "pszemraj/Phi-3-small-8k-prune6", trust_remote_code=True
)

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

microsoft/Phi-3-small-8k-instruct

Configuration

The following YAML configuration was used to produce this model:

dtype: bfloat16
merge_method: passthrough
slices:
- sources:
  - layer_range: [0, 25]
    model: microsoft/Phi-3-small-8k-instruct
- sources:
  - layer_range: [31, 32]
    model: microsoft/Phi-3-small-8k-instruct

Downloads last month: 20

Safetensors

Model size

6B params

Tensor type

BF16

Model tree for pszemraj/Phi-3-small-8k-prune6

Base model

microsoft/Phi-3-small-8k-instruct

Finetuned

(23)

this model