---
license: gpl-3.0
---

# Testing model
You can test the model on https://huggingface.co/spaces/nvomai/nvom-phi-3.5-mini-3b

q4_k_m to more optimize!

Phi 3.5 mini by Microsoft and Optimized by NVOM.ai

Benchmarks:

                           NVOM.ai AIBench SmartBench       Specials (Based on NVOM.ai bench)
    NVOM Preview 4b        7.91    10.12   9.49             Speed, Smart, Quanatisation
    Phi 3.5 mini           5.31    4.95    6.31             Smart, Speed
    Gemma 2 9b (original)  3.76    2.11    2.93             Quanatisation, Smart



1-4 -  bad

5-8 -  normal

8-12 - best


### Loading the model locally
After obtaining the Phi-3.5-mini-instruct model checkpoint, users can use this sample code for inference.

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

torch.random.manual_seed(0)

model = AutoModelForCausalLM.from_pretrained(
    "nvomai/nvom-preview-4b", 
    device_map="cuda", 
    torch_dtype="auto", 
    trust_remote_code=True, 
)
tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3.5-mini-instruct")

messages = [
    {"role": "system", "content": "You are a helpful AI assistant."},
    {"role": "user", "content": "Can you provide ways to eat combinations of bananas and dragonfruits?"},
    {"role": "assistant", "content": "Sure! Here are some ways to eat bananas and dragonfruits together: 1. Banana and dragonfruit smoothie: Blend bananas and dragonfruits together with some milk and honey. 2. Banana and dragonfruit salad: Mix sliced bananas and dragonfruits together with some lemon juice and honey."},
    {"role": "user", "content": "What about solving an 2x + 3 = 7 equation?"},
]

pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
)

generation_args = {
    "max_new_tokens": 500,
    "return_full_text": False,
    "temperature": 0.0,
    "do_sample": False,
}

output = pipe(messages, **generation_args)
print(output[0]['generated_text'])
```