|
|
from .base_prompter import BasePrompter |
|
|
from ..models.model_manager import ModelManager |
|
|
from ..models import HunyuanDiTCLIPTextEncoder, HunyuanDiTT5TextEncoder |
|
|
from transformers import BertTokenizer, AutoTokenizer |
|
|
import warnings, os |
|
|
|
|
|
|
|
|
class HunyuanDiTPrompter(BasePrompter): |
|
|
def __init__( |
|
|
self, |
|
|
tokenizer_path=None, |
|
|
tokenizer_t5_path=None |
|
|
): |
|
|
if tokenizer_path is None: |
|
|
base_path = os.path.dirname(os.path.dirname(__file__)) |
|
|
tokenizer_path = os.path.join(base_path, "tokenizer_configs/hunyuan_dit/tokenizer") |
|
|
if tokenizer_t5_path is None: |
|
|
base_path = os.path.dirname(os.path.dirname(__file__)) |
|
|
tokenizer_t5_path = os.path.join(base_path, "tokenizer_configs/hunyuan_dit/tokenizer_t5") |
|
|
super().__init__() |
|
|
self.tokenizer = BertTokenizer.from_pretrained(tokenizer_path) |
|
|
with warnings.catch_warnings(): |
|
|
warnings.simplefilter("ignore") |
|
|
self.tokenizer_t5 = AutoTokenizer.from_pretrained(tokenizer_t5_path) |
|
|
self.text_encoder: HunyuanDiTCLIPTextEncoder = None |
|
|
self.text_encoder_t5: HunyuanDiTT5TextEncoder = None |
|
|
|
|
|
|
|
|
def fetch_models(self, text_encoder: HunyuanDiTCLIPTextEncoder = None, text_encoder_t5: HunyuanDiTT5TextEncoder = None): |
|
|
self.text_encoder = text_encoder |
|
|
self.text_encoder_t5 = text_encoder_t5 |
|
|
|
|
|
|
|
|
def encode_prompt_using_signle_model(self, prompt, text_encoder, tokenizer, max_length, clip_skip, device): |
|
|
text_inputs = tokenizer( |
|
|
prompt, |
|
|
padding="max_length", |
|
|
max_length=max_length, |
|
|
truncation=True, |
|
|
return_attention_mask=True, |
|
|
return_tensors="pt", |
|
|
) |
|
|
text_input_ids = text_inputs.input_ids |
|
|
attention_mask = text_inputs.attention_mask.to(device) |
|
|
prompt_embeds = text_encoder( |
|
|
text_input_ids.to(device), |
|
|
attention_mask=attention_mask, |
|
|
clip_skip=clip_skip |
|
|
) |
|
|
return prompt_embeds, attention_mask |
|
|
|
|
|
|
|
|
def encode_prompt( |
|
|
self, |
|
|
prompt, |
|
|
clip_skip=1, |
|
|
clip_skip_2=1, |
|
|
positive=True, |
|
|
device="cuda" |
|
|
): |
|
|
prompt = self.process_prompt(prompt, positive=positive) |
|
|
|
|
|
|
|
|
prompt_emb, attention_mask = self.encode_prompt_using_signle_model(prompt, self.text_encoder, self.tokenizer, self.tokenizer.model_max_length, clip_skip, device) |
|
|
|
|
|
|
|
|
prompt_emb_t5, attention_mask_t5 = self.encode_prompt_using_signle_model(prompt, self.text_encoder_t5, self.tokenizer_t5, self.tokenizer_t5.model_max_length, clip_skip_2, device) |
|
|
|
|
|
return prompt_emb, attention_mask, prompt_emb_t5, attention_mask_t5 |
|
|
|
