File size: 9,738 Bytes
714cf46 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 | """
Data-driven HuggingFace upload script for all FastPLMs models.
Runs weight conversion scripts for each family, then uploads
modeling code, embedding_mixin, entrypoint_setup, readmes, and licenses
to each HF repo.
Usage:
py -m update_HF
py -m update_HF --hf_token YOUR_TOKEN
py -m update_HF --families esm2 dplm
py -m update_HF --skip-weights
py -m update_HF --files-only
"""
import argparse
import platform
import subprocess
from huggingface_hub import HfApi, login
MODEL_REGISTRY = [
{
"family": "e1",
"repo_ids": [
"Synthyra/Profluent-E1-150M",
"Synthyra/Profluent-E1-300M",
"Synthyra/Profluent-E1-600M",
],
"files": {
"e1_fastplms/modeling_e1.py": "modeling_e1.py",
"e1_fastplms/tokenizer.json": "tokenizer.json",
},
"readme_map": {
"Synthyra/Profluent-E1-150M": "readmes/e1_readme.md",
"Synthyra/Profluent-E1-300M": "readmes/e1_readme.md",
"Synthyra/Profluent-E1-600M": "readmes/e1_readme.md",
},
"license": "LICENSE",
"weight_module": "e1_fastplms.get_e1_weights",
},
{
"family": "esmplusplus",
"repo_ids": [
"Synthyra/ESMplusplus_small",
"Synthyra/ESMplusplus_large",
],
"files": {
"esm_plusplus/modeling_esm_plusplus.py": "modeling_esm_plusplus.py",
},
"readme_map": {
"Synthyra/ESMplusplus_small": "readmes/esm_plusplus_small_readme.md",
"Synthyra/ESMplusplus_large": "readmes/esm_plusplus_large_readme.md",
},
"license": "LICENSE",
"weight_module": "esm_plusplus.get_esmc_weights",
},
{
"family": "esm2",
"repo_ids": [
"Synthyra/ESM2-8M",
"Synthyra/ESM2-35M",
"Synthyra/ESM2-150M",
"Synthyra/ESM2-650M",
"Synthyra/ESM2-3B",
"Synthyra/FastESM2_650",
],
"files": {
"esm2/modeling_fastesm.py": "modeling_fastesm.py",
},
"readme_map": {
"Synthyra/ESM2-8M": "readmes/fastesm2_readme.md",
"Synthyra/ESM2-35M": "readmes/fastesm2_readme.md",
"Synthyra/ESM2-150M": "readmes/fastesm2_readme.md",
"Synthyra/ESM2-650M": "readmes/fastesm2_readme.md",
"Synthyra/ESM2-3B": "readmes/fastesm2_readme.md",
"Synthyra/FastESM2_650": "readmes/fastesm_650_readme.md",
},
"license": "LICENSE",
"weight_module": "esm2.get_esm2_weights",
},
{
"family": "dplm",
"repo_ids": [
"Synthyra/DPLM-150M",
"Synthyra/DPLM-650M",
"Synthyra/DPLM-3B",
],
"files": {
"dplm_fastplms/modeling_dplm.py": "modeling_dplm.py",
},
"readme_map": {
"Synthyra/DPLM-150M": "readmes/dplm_readme.md",
"Synthyra/DPLM-650M": "readmes/dplm_readme.md",
"Synthyra/DPLM-3B": "readmes/dplm_readme.md",
},
"license": "LICENSE",
"weight_module": "dplm_fastplms.get_dplm_weights",
},
{
"family": "dplm2",
"repo_ids": [
"Synthyra/DPLM2-150M",
"Synthyra/DPLM2-650M",
"Synthyra/DPLM2-3B",
],
"files": {
"dplm2_fastplms/modeling_dplm2.py": "modeling_dplm2.py",
},
"readme_map": {
"Synthyra/DPLM2-150M": "readmes/dplm2_readme.md",
"Synthyra/DPLM2-650M": "readmes/dplm2_readme.md",
"Synthyra/DPLM2-3B": "readmes/dplm2_readme.md",
},
"license": "LICENSE",
"weight_module": "dplm2_fastplms.get_dplm2_weights",
},
{
"family": "ankh",
"repo_ids": [
"Synthyra/ANKH_base",
"Synthyra/ANKH_large",
"Synthyra/ANKH2_large",
],
"files": {},
"readme_map": {},
"license": "LICENSE",
"weight_module": None,
},
{
"family": "boltz",
"repo_ids": [
"Synthyra/Boltz2",
],
"files": {
"boltz_fastplms/modeling_boltz2.py": "modeling_boltz2.py",
"boltz_fastplms/__init__.py": "__init__.py",
"boltz_fastplms/minimal_featurizer.py": "minimal_featurizer.py",
"boltz_fastplms/minimal_structures.py": "minimal_structures.py",
"boltz_fastplms/cif_writer.py": "cif_writer.py",
"boltz_fastplms/vb_const.py": "vb_const.py",
"boltz_fastplms/vb_layers_attention.py": "vb_layers_attention.py",
"boltz_fastplms/vb_layers_attentionv2.py": "vb_layers_attentionv2.py",
"boltz_fastplms/vb_layers_confidence_utils.py": "vb_layers_confidence_utils.py",
"boltz_fastplms/vb_layers_dropout.py": "vb_layers_dropout.py",
"boltz_fastplms/vb_layers_initialize.py": "vb_layers_initialize.py",
"boltz_fastplms/vb_layers_outer_product_mean.py": "vb_layers_outer_product_mean.py",
"boltz_fastplms/vb_layers_pair_averaging.py": "vb_layers_pair_averaging.py",
"boltz_fastplms/vb_layers_pairformer.py": "vb_layers_pairformer.py",
"boltz_fastplms/vb_layers_transition.py": "vb_layers_transition.py",
"boltz_fastplms/vb_layers_triangular_mult.py": "vb_layers_triangular_mult.py",
"boltz_fastplms/vb_loss_diffusionv2.py": "vb_loss_diffusionv2.py",
"boltz_fastplms/vb_modules_confidencev2.py": "vb_modules_confidencev2.py",
"boltz_fastplms/vb_modules_diffusion_conditioning.py": "vb_modules_diffusion_conditioning.py",
"boltz_fastplms/vb_modules_diffusionv2.py": "vb_modules_diffusionv2.py",
"boltz_fastplms/vb_modules_encodersv2.py": "vb_modules_encodersv2.py",
"boltz_fastplms/vb_modules_transformersv2.py": "vb_modules_transformersv2.py",
"boltz_fastplms/vb_modules_trunkv2.py": "vb_modules_trunkv2.py",
"boltz_fastplms/vb_modules_utils.py": "vb_modules_utils.py",
"boltz_fastplms/vb_potentials_potentials.py": "vb_potentials_potentials.py",
"boltz_fastplms/vb_potentials_schedules.py": "vb_potentials_schedules.py",
"boltz_fastplms/vb_tri_attn_attention.py": "vb_tri_attn_attention.py",
"boltz_fastplms/vb_tri_attn_primitives.py": "vb_tri_attn_primitives.py",
"boltz_fastplms/vb_tri_attn_utils.py": "vb_tri_attn_utils.py",
},
"readme_map": {
"Synthyra/Boltz2": "readmes/boltz2_readme.md",
},
"license": "LICENSE",
"weight_module": "boltz_fastplms.get_boltz2_weights",
},
]
SHARED_FILES = {
"embedding_mixin.py": "embedding_mixin.py",
"entrypoint_setup.py": "entrypoint_setup.py",
}
def _run_weight_scripts(
families: list[str] | None, hf_token: str | None, skip_weights: bool
) -> None:
python_cmd = "python" if platform.system().lower() == "linux" else "py"
for entry in MODEL_REGISTRY:
if families is not None and entry["family"] not in families:
continue
module = entry["weight_module"]
if module is None:
continue
command = [python_cmd, "-m", module]
if hf_token is not None:
command.extend(["--hf_token", hf_token])
if skip_weights:
command.append("--skip-weights")
print(f"Running: {' '.join(command)}")
subprocess.run(command, check=True)
def _upload_files(api: HfApi, families: list[str] | None) -> None:
for entry in MODEL_REGISTRY:
if families is not None and entry["family"] not in families:
continue
for repo_id in entry["repo_ids"]:
print(f"\nUploading to {repo_id}")
for local_path, repo_path in entry["files"].items():
api.upload_file(
path_or_fileobj=local_path,
path_in_repo=repo_path,
repo_id=repo_id,
repo_type="model",
)
for local_path, repo_path in SHARED_FILES.items():
api.upload_file(
path_or_fileobj=local_path,
path_in_repo=repo_path,
repo_id=repo_id,
repo_type="model",
)
if entry["license"]:
api.upload_file(
path_or_fileobj=entry["license"],
path_in_repo="LICENSE",
repo_id=repo_id,
repo_type="model",
)
readme_path = entry["readme_map"].get(repo_id)
if readme_path:
api.upload_file(
path_or_fileobj=readme_path,
path_in_repo="README.md",
repo_id=repo_id,
repo_type="model",
)
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Upload FastPLMs models to HuggingFace")
parser.add_argument("--hf_token", type=str, default=None)
parser.add_argument("--families", nargs="+", default=None)
parser.add_argument(
"--skip-weights",
action="store_true",
help="Run weight scripts without downloading/pushing model weights",
)
parser.add_argument("--files-only", action="store_true", help="Only upload files, skip weight conversion")
args = parser.parse_args()
if args.hf_token:
login(token=args.hf_token)
if not args.files_only:
_run_weight_scripts(args.families, args.hf_token, args.skip_weights)
api = HfApi()
_upload_files(api, args.families)
print("\nDone.")
|