2024-12-26 06:01:57 -05:00
|
|
|
import os
|
|
|
|
import json
|
|
|
|
from argparse import ArgumentParser
|
|
|
|
from glob import glob
|
2025-01-29 00:07:09 -05:00
|
|
|
from typing import Dict, Any
|
2024-12-26 06:01:57 -05:00
|
|
|
from tqdm import tqdm
|
|
|
|
|
|
|
|
import torch
|
|
|
|
from safetensors.torch import load_file, save_file
|
|
|
|
|
|
|
|
from kernel import weight_dequant
|
|
|
|
|
2025-01-05 13:18:18 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
class WeightConverter:
|
|
|
|
def __init__(self, fp8_path: str, bf16_path: str):
|
|
|
|
"""
|
|
|
|
Initialize the weight converter with input and output paths.
|
2025-01-05 13:18:18 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
Args:
|
|
|
|
fp8_path (str): Path to the directory containing FP8 weights
|
|
|
|
bf16_path (str): Path to save the converted BF16 weights
|
|
|
|
"""
|
|
|
|
self.fp8_path = fp8_path
|
|
|
|
self.bf16_path = bf16_path
|
|
|
|
self.loaded_files: Dict[str, Dict[str, torch.Tensor]] = {}
|
|
|
|
self.fp8_weight_names: list = []
|
|
|
|
self.weight_map: Dict[str, str] = self._load_model_index()
|
2025-01-05 13:18:18 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
def _load_model_index(self) -> Dict[str, str]:
|
|
|
|
"""
|
|
|
|
Load the model index file.
|
2025-01-05 13:18:18 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
Returns:
|
|
|
|
Dict[str, str]: Weight mapping from the index file
|
|
|
|
"""
|
|
|
|
model_index_file = os.path.join(self.fp8_path, "model.safetensors.index.json")
|
|
|
|
with open(model_index_file, "r") as f:
|
|
|
|
return json.load(f)["weight_map"]
|
|
|
|
|
|
|
|
def _get_tensor(self, tensor_name: str) -> torch.Tensor:
|
2025-01-05 13:18:18 -05:00
|
|
|
"""
|
2025-01-29 00:07:09 -05:00
|
|
|
Get a tensor from cache or load it from disk.
|
2025-01-05 13:18:18 -05:00
|
|
|
|
|
|
|
Args:
|
2025-01-29 00:07:09 -05:00
|
|
|
tensor_name (str): Name of the tensor to retrieve
|
2025-01-05 13:18:18 -05:00
|
|
|
|
|
|
|
Returns:
|
2025-01-29 00:07:09 -05:00
|
|
|
torch.Tensor: The requested tensor
|
2025-01-05 13:18:18 -05:00
|
|
|
|
|
|
|
Raises:
|
2025-01-29 00:07:09 -05:00
|
|
|
KeyError: If tensor doesn't exist in the safetensor file
|
|
|
|
"""
|
|
|
|
file_name = self.weight_map[tensor_name]
|
|
|
|
if file_name not in self.loaded_files:
|
|
|
|
file_path = os.path.join(self.fp8_path, file_name)
|
|
|
|
self.loaded_files[file_name] = load_file(file_path, device="cuda")
|
|
|
|
return self.loaded_files[file_name][tensor_name]
|
|
|
|
|
|
|
|
def _manage_memory(self):
|
|
|
|
"""
|
|
|
|
Keep only the 2 most recently used files in memory.
|
|
|
|
"""
|
|
|
|
if len(self.loaded_files) > 2:
|
|
|
|
oldest_file = next(iter(self.loaded_files))
|
|
|
|
del self.loaded_files[oldest_file]
|
|
|
|
torch.cuda.empty_cache()
|
|
|
|
|
|
|
|
def _process_weight(self, weight_name: str, weight: torch.Tensor) -> torch.Tensor:
|
|
|
|
"""
|
|
|
|
Process a single weight tensor.
|
|
|
|
|
|
|
|
Args:
|
|
|
|
weight_name (str): Name of the weight tensor
|
|
|
|
weight (torch.Tensor): The weight tensor to process
|
|
|
|
|
|
|
|
Returns:
|
|
|
|
torch.Tensor: Processed weight tensor
|
2025-01-05 13:18:18 -05:00
|
|
|
"""
|
2025-01-29 00:07:09 -05:00
|
|
|
if weight_name.endswith("_scale_inv"):
|
|
|
|
return None
|
2024-12-26 20:34:38 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
if weight.element_size() == 1: # FP8 weight
|
|
|
|
scale_inv_name = f"{weight_name}_scale_inv"
|
|
|
|
try:
|
|
|
|
scale_inv = self._get_tensor(scale_inv_name)
|
|
|
|
self.fp8_weight_names.append(weight_name)
|
|
|
|
return weight_dequant(weight, scale_inv)
|
|
|
|
except KeyError:
|
|
|
|
print(f"Warning: Missing scale_inv tensor for {weight_name}, skipping conversion")
|
|
|
|
return weight
|
|
|
|
return weight
|
|
|
|
|
|
|
|
def _save_model_index(self):
|
|
|
|
"""
|
|
|
|
Save the updated model index file.
|
|
|
|
"""
|
|
|
|
new_model_index_file = os.path.join(self.bf16_path, "model.safetensors.index.json")
|
|
|
|
for weight_name in self.fp8_weight_names:
|
|
|
|
scale_inv_name = f"{weight_name}_scale_inv"
|
|
|
|
if scale_inv_name in self.weight_map:
|
|
|
|
self.weight_map.pop(scale_inv_name)
|
2024-12-26 20:34:38 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
with open(new_model_index_file, "w") as f:
|
|
|
|
json.dump({"metadata": {}, "weight_map": self.weight_map}, f, indent=2)
|
|
|
|
|
|
|
|
def convert(self):
|
|
|
|
"""
|
|
|
|
Convert FP8 weights to BF16 format.
|
|
|
|
"""
|
|
|
|
torch.set_default_dtype(torch.bfloat16)
|
|
|
|
os.makedirs(self.bf16_path, exist_ok=True)
|
2024-12-26 20:34:38 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
safetensor_files = sorted(glob(os.path.join(self.fp8_path, "*.safetensors")))
|
|
|
|
|
|
|
|
for safetensor_file in tqdm(safetensor_files):
|
|
|
|
file_name = os.path.basename(safetensor_file)
|
|
|
|
current_state_dict = load_file(safetensor_file, device="cuda")
|
|
|
|
self.loaded_files[file_name] = current_state_dict
|
|
|
|
|
|
|
|
new_state_dict = {}
|
|
|
|
for weight_name, weight in current_state_dict.items():
|
|
|
|
processed_weight = self._process_weight(weight_name, weight)
|
|
|
|
if processed_weight is not None:
|
|
|
|
new_state_dict[weight_name] = processed_weight
|
|
|
|
|
|
|
|
new_safetensor_file = os.path.join(self.bf16_path, file_name)
|
|
|
|
save_file(new_state_dict, new_safetensor_file)
|
|
|
|
|
|
|
|
self._manage_memory()
|
2024-12-26 06:01:57 -05:00
|
|
|
|
2025-01-29 00:07:09 -05:00
|
|
|
self._save_model_index()
|
|
|
|
|
|
|
|
|
|
|
|
def main(fp8_path: str, bf16_path: str):
|
|
|
|
"""
|
|
|
|
Main function to convert FP8 weights to BF16.
|
|
|
|
|
|
|
|
Args:
|
|
|
|
fp8_path (str): Input directory containing FP8 weights
|
|
|
|
bf16_path (str): Output directory for BF16 weights
|
|
|
|
"""
|
|
|
|
converter = WeightConverter(fp8_path, bf16_path)
|
|
|
|
converter.convert()
|
|
|
|
|
2024-12-26 06:01:57 -05:00
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
parser = ArgumentParser()
|
|
|
|
parser.add_argument("--input-fp8-hf-path", type=str, required=True)
|
|
|
|
parser.add_argument("--output-bf16-hf-path", type=str, required=True)
|
|
|
|
args = parser.parse_args()
|
2025-01-29 00:07:09 -05:00
|
|
|
main(args.input_fp8_hf_path, args.output_bf16_hf_path)
|