lxuechen/phi-2-sft

Model Summary

phi-2-sft is an instruction-tuned model from the original Phi-2 on OpenAssistant instruction-following data.

The purpose of the experiment is to understand the quality of the pre-trained Phi-2 model. The good news is that phi-2-sft can follow open-ended user instructions well.

Decoding

Format your prompt as

"""### Human: {instruction}

### Assistant:"""

where instruction is your query.

Here's a full-fledged example:

import torch
import transformers

model_name_or_path = "lxuechen/phi-2-sft"

model: transformers.PreTrainedModel = transformers.AutoModelForCausalLM.from_pretrained(
    model_name_or_path,
    low_cpu_mem_usage=True,
    device_map="auto",
    trust_remote_code=True,
    torch_dtype=torch.float16
)
tokenizer = transformers.AutoTokenizer.from_pretrained(model_name_or_path)

input_text = "### Human: Give me a good recipe for a chinese dish\n\n### Assistant:"

outputs = model.generate(
    tokenizer(input_text, return_tensors="pt").to(model.device),
    max_length=1024,
    temperature=0.7,
    top_p=0.9,
    do_sample=True,
    pad_token_id=tokenizer.pad_token_id,
    eos_token_id=tokenizer.eos_token_id,
)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training

The model was fine-tuned on a sanitized version of the single-turn OASST dataset linked above.

Hyperparameters:

• learning rate: 3% linear warmup, with a peak of 2e-5 and cosine decay
• epochs: 2
• batch size: 64
• context length: 512

Limitations of phi-2-sft

• Generate Inaccurate Code and Facts: The model may produce 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: Majority of Phi-2 training data is based in Python and use common packages such as "typing, math, random, collections, datetime, itertools". If the model generates Python scripts that utilize other 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 languages might pose challenges to its comprehension, leading to potential misinterpretations or errors in response.

• Potential Societal Biases: Phi-2 is not entirely free from societal biases despite efforts in assuring trainig data safety. 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 being 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.

• Verbosity: Phi-2 being a base model often produces irrelevant or extra text and responses following its first answer to user prompts within a single turn. This is due to its training dataset being primarily textbooks, which results in textbook-like responses.