Model Summary
The language model Phi-1.5 is a Transformer with
1.3 billion parameters. It was trained using the same data sources as
phi-1, augmented with a new data source that consists of various NLP synthetic texts. When assessed against benchmarks testing common sense, language understanding, and logical reasoning, Phi-1.5 demonstrates a nearly state-of-the-art performance among models with less than 10 billion parameters.
We did not fine-tune Phi-1.5 either for instruction following or through reinforcement learning from human feedback. The intention behind crafting this open-source model is to provide the research community with a non-restricted small model to explore vital safety challenges, such as reducing toxicity, understanding societal biases, enhancing controllability, and more.
For a safer model release, we exclude generic web-crawl data sources such as common-crawl from the training. This strategy prevents direct exposure to potentially harmful online content, enhancing the model's safety without RLHF. However, the model is still vulnerable to generating harmful content. We hope the model can help the research community to further study the safety of language models.
Phi-1.5 can write poems, draft emails, create stories, summarize texts, write Python code (such as downloading a Hugging Face transformer model), etc.
How to Use
Phi-1.5 has been integrated in the transformers version 4.37.0. If you are using a lower version, ensure that you are doing the following:
- When loading the model, ensure that
trust_remote_code=True is passed as an argument of the from_pretrained() function.
The current transformers version can be verified with: pip list | grep transformers.
Intended Uses
Given the nature of the training data, Phi-1.5 is best suited for prompts using the QA format, the chat format, and the code format. Note that Phi-1.5, being a base model, often produces irrelevant text following the main answer. In the following example, we've truncated the answer for illustrative purposes only.
QA Format:
1Write a detailed analogy between mathematics and a lighthouse.
2
3Answer: Mathematics is like a lighthouse, guiding us through the vast ocean of numbers and calculations. Just as a lighthouse illuminates the darkness, mathematics provides us with a clear path to navigate through complex problems. It helps us make sense of the world around us, just like a lighthouse helps ships find their way home.
where the model generates the text after "Answer:".
Chat Format:
1Alice: I don't know why, I'm struggling to maintain focus while studying. Any suggestions?
2
3Bob: Have you tried using a timer? It can help you stay on track and avoid distractions.
4
5Alice: That's a good idea. I'll give it a try.
6
7Charlie: Another thing that can help is to break up your study sessions into smaller chunks. It's easier to concentrate on one thing at a time.
8
9Alice: That makes sense. I'll try that too.
10
11Bob: And don't forget to take breaks! It's important to give your brain a rest so you can come back to your studies with a fresh perspective.
12
13Alice: Thanks for the advice, guys. I feel more motivated now.
14
15Charlie: No problem, Alice. We're all in this together.
16
17Bob: Yeah, and remember that it's okay to ask for help if you need it. We're here to support each other.
where the model generates the text after the first "Bob:".
Code Format:
1def print_prime(n):
2 """
3 Print all primes between 1 and n
4 """
5 primes = []
6 for num in range(2, n+1):
7 is_prime = True
8 for i in range(2, int(math.sqrt(num))+1):
9 if num % i == 0:
10 is_prime = False
11 break
12 if is_prime:
13 primes.append(num)
14 print(primes)
where the model generates the text after the comments.
Notes:
-
Phi-1.5-generated text/code should be treated as a starting point rather than a definitive solution for potential use cases. Users should be cautious when employing these models in their applications.
-
Phi-1.5 has not been tested to ensure that it performs adequately for any production-level application. Please refer to the limitation sections of this document for more details.
-
If you are using transformers<4.37.0, always load the model with trust_remote_code=True to prevent side-effects.
Sample Code
1import torch
2from transformers import AutoModelForCausalLM, AutoTokenizer
3
4torch.set_default_device("cuda")
5
6model = AutoModelForCausalLM.from_pretrained("microsoft/phi-1_5", torch_dtype="auto", trust_remote_code=True)
7tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-1_5", trust_remote_code=True)
8
9inputs = tokenizer('''def print_prime(n):
10 """
11 Print all primes between 1 and n
12 """''', return_tensors="pt", return_attention_mask=False)
13
14outputs = model.generate(**inputs, max_length=200)
15text = tokenizer.batch_decode(outputs)[0]
16print(text)
Limitations of Phi-1.5
-
Generate Inaccurate Code and Facts: The model often produces incorrect code snippets and statements. Users should treat these outputs as suggestions or starting points, not as definitive or accurate solutions.
-
Limited Scope for code: If the model generates Python scripts that utilize uncommon packages or scripts in other languages, we strongly recommend users manually verify all API uses.
-
Unreliable Responses to Instruction: The model has not undergone instruction fine-tuning. As a result, it may struggle or fail to adhere to intricate or nuanced instructions provided by users.
-
Language Limitations: The model is primarily designed to understand standard English. Informal English, slang, or any other language outside of English might pose challenges to its comprehension, leading to potential misinterpretations or errors in response.
-
Potential Societal Biases: Regardless of the safe data used for its training, the model is not entirely free from societal biases. There's a possibility it may generate content that mirrors these societal biases, particularly if prompted or instructed to do so. We urge users to be aware of this and to exercise caution and critical thinking when interpreting model outputs.
-
Toxicity: Despite that the model is trained with carefully selected data, the model can still produce harmful content if explicitly prompted or instructed to do so. We chose to release the model for research purposes only -- We hope to help the open-source community develop the most effective ways to reduce the toxicity of a model directly after pretraining.
Training
Model
- Architecture: a Transformer-based model with next-word prediction objective
- Dataset size: 30B tokens
- Training tokens: 150B tokens
- Precision: fp16
- GPUs: 32xA100-40G
- Training time: 8 days
Software
License
The model is licensed under the
MIT license.
Citation
You can find the paper at
https://arxiv.org/abs/2309.05463. Please cite as:
1@article{textbooks2,
2 title={Textbooks Are All You Need II: \textbf{phi-1.5} technical report},
3 author={Li, Yuanzhi and Bubeck, S{\'e}bastien and Eldan, Ronen and Del Giorno, Allie and Gunasekar, Suriya and Lee, Yin Tat},
4 journal={arXiv preprint arXiv:2309.05463},
5 year={2023}
6}
Trademarks
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow
Microsoft’s Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party’s policies.
