PoCs a utility to get the totsl parameters count of a diffusers pipeline without hefty downloads.

get_params_diffusers_diffusers.py

from huggingface_hub import (
    hf_hub_download,
    list_files_info,
    parse_safetensors_file_metadata,
)
import json
import argparse


def get_files_info(args):
    files_info = list_files_info(args.repo_id, expand=True)
    files_info = list(files_info)
    return files_info


def get_model_index(args):
    model_index_path = hf_hub_download(
        repo_id=args.repo_id, filename="model_index.json"
    )

    with open(model_index_path) as f:
        model_index = json.load(f)

    return model_index


def get_params(args):
    model_index = get_model_index(args)
    components = [k for k in model_index.keys() if isinstance(model_index[k], list)]

    repo_files_info = get_files_info(args)

    total_params = 0
    for file in repo_files_info:
        if (
            len(file.path.split("/")[-1].split(".")) == 2
            and "safetensors" in file.path
            and any(k in file.path for k in components)
        ):
            total_params += parse_safetensors_file_metadata(
                repo_id=args.repo_id, filename=file.path
            ).parameter_count["F32"]

    return total_params


if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="Simple example of a training script.")
    parser.add_argument(
        "--repo_id",
        type=str,
        default="runwayml/stable-diffusion-v1-5",
        help="Repo id to count the params for.",
    )
    args = parser.parse_args()

    params = get_params(args)

    print(f"{args.repo_id} has {params / 1e9} billion parameters.")

model_size_poc_diffusers.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "eb50220e-ea67-45d4-be7b-01f824118dcd",
   "metadata": {},
   "outputs": [],
   "source": [
    "from huggingface_hub import hf_hub_download, list_files_info\n",
    "import safetensors\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f59a730d-361c-4a85-a49b-b989a3198343",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/sayakpaul/miniconda3/envs/diffusers/lib/python3.9/site-packages/huggingface_hub/utils/_deprecation.py:131: FutureWarning: 'list_files_info' (from 'huggingface_hub.hf_api') is deprecated and will be removed from version '0.23'. Use `list_repo_tree` and `get_paths_info` instead.\n",
      "  warnings.warn(warning_message, FutureWarning)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[RepoFile(path='.gitattributes', size=1548, blob_id='55d2855c5be698e0572b9f42af95f06bfd5fb002', lfs=None, last_commit={'oid': 'd8c8d262a38ad3da5e673f048db73a504f2a92a9', 'title': 'Upload v1-5-pruned.ckpt', 'date': datetime.datetime(2022, 10, 20, 12, 18, 52, tzinfo=datetime.timezone.utc)}, security={'safe': True, 'av_scan': {'virusFound': False, 'virusNames': None}, 'pickle_import_scan': None}),\n",
       " RepoFile(path='README.md', size=14462, blob_id='103bad8a83037abcdb9d24f2b21eba8792b33a5d', lfs=None, last_commit={'oid': '889b629140e71758e1e0006e355c331a5744b4bf', 'title': 'Update README.md', 'date': datetime.datetime(2022, 12, 19, 15, 29, 28, tzinfo=datetime.timezone.utc)}, security={'safe': True, 'av_scan': {'virusFound': False, 'virusNames': None}, 'pickle_import_scan': None}),\n",
       " RepoFile(path='feature_extractor/preprocessor_config.json', size=342, blob_id='5294955ff7801083f720b34b55d0f1f51313c5c5', lfs=None, last_commit={'oid': '7621c1d34cb8951ae5277a5d1c07c431e46abb48', 'title': 'add diffusers weights', 'date': datetime.datetime(2022, 10, 20, 9, 30, 42, tzinfo=datetime.timezone.utc)}, security={'safe': True, 'av_scan': {'virusFound': False, 'virusNames': None}, 'pickle_import_scan': None}),\n",
       " RepoFile(path='model_index.json', size=541, blob_id='daf7e2e2dfc64fb437a2b44525667111b00cb9fc', lfs=None, last_commit={'oid': 'aa9ba505e1973ae5cd05f5aedd345178f52f8e6a', 'title': 'Fix deprecation warning by changing `CLIPFeatureExtractor` to `CLIPImageProcessor`. (#124)', 'date': datetime.datetime(2023, 5, 5, 10, 38, 18, tzinfo=datetime.timezone.utc)}, security={'safe': True, 'av_scan': {'virusFound': False, 'virusNames': None}, 'pickle_import_scan': None})]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "repo_id = \"runwayml/stable-diffusion-v1-5\"\n",
    "\n",
    "files_info = list_files_info(repo_id, expand=True)\n",
    "files_info = list(files_info)\n",
    "files_info[:4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "49f21f4a-d9f5-46f5-ad90-89e18cc4eb50",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['feature_extractor',\n",
       " 'safety_checker',\n",
       " 'scheduler',\n",
       " 'text_encoder',\n",
       " 'tokenizer',\n",
       " 'unet',\n",
       " 'vae']"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_index_path = hf_hub_download(repo_id=repo_id, filename=\"model_index.json\")\n",
    "\n",
    "with open(model_index_path) as f:\n",
    "    model_index = json.load(f)\n",
    "\n",
    "components = [k for k in model_index.keys() if isinstance(model_index[k], list)]\n",
    "components"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0b3f9f8c-50ef-4099-a7a8-e2db07961a03",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total params: 1.370216895 billion.\n"
     ]
    }
   ],
   "source": [
    "from huggingface_hub import parse_safetensors_file_metadata\n",
    "total_params = 0\n",
    "\n",
    "for file in files_info:\n",
    "    if (\n",
    "        len(file.path.split(\"/\")[-1].split(\".\")) == 2\n",
    "        and \"safetensors\" in file.path\n",
    "        and any(k in file.path for k in components)\n",
    "    ):\n",
    "        total_params += parse_safetensors_file_metadata(\n",
    "            repo_id=repo_id, filename=file.path\n",
    "        ).parameter_count[\"F32\"]\n",
    "\n",
    "print(f\"Total params: {total_params / 1e9} billion.\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.17"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}

Some results:

runwayml/stable-diffusion-v1-5

python get_params_diffusers_diffusers.py

runwayml/stable-diffusion-v1-5 has 1.370216895 billion parameters.

stabilityai/stable-diffusion-xl-base-1.0

python get_params_diffusers_diffusers.py --repo_id=stabilityai/stable-diffusion-xl-base-1.0

stabilityai/stable-diffusion-xl-base-1.0 has 3.55249173 billion parameters.

stabilityai/stable-video-diffusion-img2vid-xt

python get_params_diffusers_diffusers.py --repo_id=stabilityai/stable-video-diffusion-img2vid-xt

stabilityai/stable-video-diffusion-img2vid-xt has 2.254442729 billion parameters.