matrix-pygmalion-bot/runpod/test/runpod-worker-transformers/model_fetcher.py

'''
RunPod | Transformer | Model Fetcher
'''

import os
import argparse

import torch
from transformers import (GPTNeoForCausalLM, GPT2Tokenizer, GPTNeoXForCausalLM,
                          GPTNeoXTokenizerFast, GPTJForCausalLM, AutoTokenizer, AutoModelForCausalLM)
from huggingface_hub import snapshot_download, hf_hub_download

def download_model(model_name):

    # --------------------------------- Neo 1.3B --------------------------------- #
    if model_name == 'gpt-neo-1.3B':
        GPTNeoForCausalLM.from_pretrained("EleutherAI/gpt-neo-1.3B")
        GPT2Tokenizer.from_pretrained("EleutherAI/gpt-neo-1.3B")

    # --------------------------------- Neo 2.7B --------------------------------- #
    elif model_name == 'gpt-neo-2.7B':
        GPTNeoForCausalLM.from_pretrained("EleutherAI/gpt-neo-2.7B", torch_dtype=torch.float16)
        GPT2Tokenizer.from_pretrained("EleutherAI/gpt-neo-2.7B")

    # ----------------------------------- NeoX ----------------------------------- #
    elif model_name == 'gpt-neox-20b':
        GPTNeoXForCausalLM.from_pretrained("EleutherAI/gpt-neox-20b").half()
        GPTNeoXTokenizerFast.from_pretrained("EleutherAI/gpt-neox-20b")

    # --------------------------------- Pygmalion -------------------------------- #
    elif model_name == 'pygmalion-6b':
#        AutoModelForCausalLM.from_pretrained("PygmalionAI/pygmalion-6b", load_in_8bit=True)
#        AutoTokenizer.from_pretrained("PygmalionAI/pygmalion-6b")
        snapshot_path = snapshot_download(repo_id="PygmalionAI/pygmalion-6b", revision="main")

    # --------------------------------- Pygmalion -------------------------------- #
    elif model_name == 'pygmalion-6b-4bit-128g':
        snapshot_path = snapshot_download(repo_id="mayaeary/pygmalion-6b-4bit-128g", revision="main")

    # --------------------------------- Pygmalion -------------------------------- #
    elif model_name == 'pygmalion-6b-gptq-4bit':
#        AutoModelForCausalLM.from_pretrained("OccamRazor/pygmalion-6b-gptq-4bit", from_pt=True)
#        AutoTokenizer.from_pretrained("OccamRazor/pygmalion-6b-gptq-4bit")
        snapshot_path = snapshot_download(repo_id="OccamRazor/pygmalion-6b-gptq-4bit", revision="main")

    # ----------------------------------- GPT-J ----------------------------------- #
    elif model_name == 'gpt-j-6b':
        GPTJForCausalLM.from_pretrained("EleutherAI/gpt-j-6B", revision="float16",
                                        torch_dtype=torch.float16)
        AutoTokenizer.from_pretrained("EleutherAI/gpt-j-6B")

    # ------------------------------ PPO Shygmalion 6B ----------------------------- #
    elif model_name == 'ppo-shygmalion-6b':
        AutoModelForCausalLM.from_pretrained("TehVenom/PPO_Shygmalion-6b", load_in_8bit=True)
        AutoTokenizer.from_pretrained("TehVenom/PPO_Shygmalion-6b")

    # ------------------------------ Dolly Shygmalion 6B ----------------------------- #
    elif model_name == 'dolly-shygmalion-6b':
        AutoModelForCausalLM.from_pretrained("TehVenom/Dolly_Shygmalion-6b", load_in_8bit=True)
        AutoTokenizer.from_pretrained("TehVenom/Dolly_Shygmalion-6b")

    # ------------------------------ Erebus 13B (NSFW) ----------------------------- #
    elif model_name == 'erebus-13b':
        AutoModelForCausalLM.from_pretrained("KoboldAI/OPT-13B-Erebus", load_in_8bit=True)
        AutoTokenizer.from_pretrained("KoboldAI/OPT-13B-Erebus")

    # --------------------------- Alpaca 13B (Quantized) -------------------------- #
    elif model_name == 'gpt4-x-alpaca-13b-native-4bit-128g':
        AutoModelForCausalLM.from_pretrained("anon8231489123/gpt4-x-alpaca-13b-native-4bit-128g")
        AutoTokenizer.from_pretrained("anon8231489123/gpt4-x-alpaca-13b-native-4bit-128g")

    # --------------------------------- Alpaca 13B -------------------------------- #
    elif model_name == 'gpt4-x-alpaca':
        AutoModelForCausalLM.from_pretrained("chavinlo/gpt4-x-alpaca", load_in_8bit=True)
        AutoTokenizer.from_pretrained("chavinlo/gpt4-x-alpaca")

    # --------------------------------- RWKV Raven 7B -------------------------------- #
    elif model_name == 'rwkv-4-raven-7b':
        snapshot_path = hf_hub_download(repo_id="BlinkDL/rwkv-4-raven", filename="RWKV-4-Raven-7B-v8-Eng-20230408-ctx4096.pth")
        hf_hub_download(repo_id="BlinkDL/Raven-RWKV-7B", filename="20B_tokenizer.json", local_dir=snapshot_path)
        #https://huggingface.co/yahma/RWKV-14b_quant/resolve/main/RWKV-4-Pile-14B-20230213-8019.pqth

    if snapshot_path:
        print(f"model downloaded to \"{snapshot_path}\"")
        os.system(f"ln -s \"{snapshot_path}\" /workdir/model")

# ---------------------------------------------------------------------------- #
#                                Parse Arguments                               #
# ---------------------------------------------------------------------------- #
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("--model_name", type=str,
                    default="gpt-neo-1.3B", help="URL of the model to download.")


if __name__ == "__main__":
    args = parser.parse_args()
    download_model(args.model_name)