weight_diff.py

#  Adapted from https://github.com/tatsu-lab/stanford_alpaca/blob/main/weight_diff.py

from typing import Optional

import fire
import torch
import tqdm
import transformers


@torch.inference_mode()
def make_diff(
    path_raw: str, path_tuned: str, path_diff: str, device="cpu",  # "cuda" or "cpu"
):
    """Make the weight diff.

    This function is given to present full transparency of how the weight diff was created.

    Run:
        python weight_diff.py make_diff --path_raw <your_path_raw> --path_tuned <your_path_tuned> --path_diff <your_path_diff>
    """
    model_tuned: transformers.PreTrainedModel = transformers.AutoModelForCausalLM.from_pretrained(
        path_tuned,
        device_map={"": torch.device(device)},
        torch_dtype=torch.float32,
        low_cpu_mem_usage=True,
    )
    model_raw: transformers.PreTrainedModel = transformers.AutoModelForCausalLM.from_pretrained(
        path_raw,
        device_map={"": torch.device(device)},
        torch_dtype=torch.float32,
        low_cpu_mem_usage=True,
    )

    tokenizer_tuned: transformers.PreTrainedTokenizer = transformers.AutoTokenizer.from_pretrained(
        path_tuned
    )
    tokenizer_raw: transformers.PreTrainedTokenizer = transformers.AutoTokenizer.from_pretrained(
        path_raw
    )
    
    tokenizer_tuned.pad_token_id = (
        0  # unk. we want this to be different from the eos token
    )
    tokenizer_tuned.padding_side = "left"  # Allow batched inference

    state_dict_tuned = model_tuned.state_dict()
    state_dict_raw = model_raw.state_dict()
    for key in tqdm.tqdm(state_dict_tuned):
        state_dict_tuned[key].add_(-state_dict_raw[key])

    model_tuned.save_pretrained(path_diff)
    tokenizer_tuned.save_pretrained(path_diff)


@torch.inference_mode()
def recover(
    path_raw,
    path_diff,
    path_tuned: Optional[str] = None,
    device="cpu",
    test_inference=True
):
    """Recover the original weights from the released weight diff.

    This function is given for you to run.

    Things to do before running this:
        1. Convert Meta's released weights into huggingface format. Follow this guide:
            https://huggingface.co/docs/transformers/main/model_doc/llama
            You may refer to https://huggingface.co/huggyllama/llama-7b if you get some trouble in the conversion. (You should only use this repository if you have been granted access to the llama model.)
        2. Make sure you cloned the released weight diff into your local machine. The weight diff is located at:
            https://huggingface.co/Dynosaur/dynosaur-llama-7b-superni
        3. Run this function with the correct paths. E.g.,
            python weight_diff.py recover --path_raw <path_to_step_1_dir> --path_diff <path_to_step_2_dir>

    Additional notes:
        - If things run too slowly, and you have an 80G GPU lying around, let GPU go brrr by setting `--device "cuda"`.
        - If you want to save the recovered weights, set `--path_tuned <your_path_tuned>`.
            Next time you can load the recovered weights directly from `<your_path_tuned>`.
    """
    model_raw: transformers.PreTrainedModel = transformers.AutoModelForCausalLM.from_pretrained(
        path_raw,
        device_map={"": torch.device(device)},
        torch_dtype=torch.float32,
        low_cpu_mem_usage=True,
    )
    model_recovered: transformers.PreTrainedModel = transformers.AutoModelForCausalLM.from_pretrained(
        path_diff,
        device_map={"": torch.device(device)},
        torch_dtype=torch.float32,
        low_cpu_mem_usage=True,
    )

    tokenizer_raw: transformers.PreTrainedTokenizer = transformers.AutoTokenizer.from_pretrained(
        path_raw
    )
    tokenizer_recovered: transformers.PreTrainedTokenizer = transformers.AutoTokenizer.from_pretrained(
        path_diff
    )

    state_dict_recovered = model_recovered.state_dict()
    state_dict_raw = model_raw.state_dict()
    for key in tqdm.tqdm(state_dict_recovered):
        state_dict_recovered[key].add_(state_dict_raw[key])

    if path_tuned is not None:
        model_recovered.save_pretrained(path_tuned)
        tokenizer_recovered.save_pretrained(path_tuned)

    if test_inference:
        input_text = (
            "Below is an instruction that describes a task. "
            "Write a response that appropriately completes the request.\r\n\r\n"
            "### Instruction:\r\nList three technologies that make life easier.\r\n\r\n### Response:"
        )
        inputs = tokenizer_recovered(input_text, return_tensors="pt")
        out = model_recovered.generate(inputs=inputs.input_ids, max_new_tokens=100)
        output_text = tokenizer_recovered.batch_decode(out, skip_special_tokens=True)[0]
        output_text = output_text[len(input_text) :]
        print(f"Input: {input_text}\nCompletion: {output_text}")

    return model_recovered, tokenizer_recovered


def main(task, **kwargs):
    globals()[task](**kwargs)


if __name__ == "__main__":
    fire.Fire(main)