remove_pooling_from_te

README.md

@@ -28,55 +28,18 @@ pip install git+https://github.com/huggingface/diffusers
 ```py
 import torch
 from diffusers import DiffusionPipeline
-
-from transformers import XLMRobertaTokenizerFast,XLMRobertaModel
-from diffusers import FlowMatchEulerDiscreteScheduler
-from diffusers.models import AutoencoderKL
-from diffusers import SanaTransformer2DModel
+#from pipeline_waifu import WaifuPipeline
 
 pipe_id = "AiArtLab/waifu-2b"
 variant = "fp16"
-# tokenizer
-tokenizer = XLMRobertaTokenizerFast.from_pretrained(
-    pipe_id,
-    subfolder="tokenizer"
-)
-
-# text_encoder
-text_encoder = XLMRobertaModel.from_pretrained(
-    pipe_id,
-    variant=variant,
-    subfolder="text_encoder",
-    add_pooling_layer=False
-).to("cuda")
-
-# scheduler
-scheduler = FlowMatchEulerDiscreteScheduler(shift=1.0)
-
-# VAE
-vae = AutoencoderKL.from_pretrained(
-    pipe_id,
-    variant=variant,
-    subfolder="vae"
-).to("cuda")
-
-# Transformer
-transformer = SanaTransformer2DModel.from_pretrained(
-    pipe_id,
-    variant=variant,
-    subfolder="transformer"
-).to("cuda")
 
 # Pipeline
 pipeline = DiffusionPipeline.from_pretrained(
     pipe_id,
-    tokenizer=tokenizer,
-    text_encoder=text_encoder,
-    vae=vae,
-    transformer=transformer,
-    trust_remote_code=True,
+    variant=variant,
+    trust_remote_code = True
 ).to("cuda")
-print(pipeline)
+#print(pipeline)
 
 prompt = 'аниме девушка, waifu, يبتسم جنسيا , sur le fond de la tour Eiffel'
 generator = torch.Generator(device="cuda").manual_seed(42)

pipeline_waifu.py

@@ -127,6 +127,8 @@ class WaifuPipeline(DiffusionPipeline):
             tokenizer=tokenizer, text_encoder=text_encoder, vae=vae, transformer=transformer, scheduler=scheduler
         )
 
+        self.text_encoder.pooler = None
+
         self.vae_scale_factor = (
             8
         )

transformer/config.json

@@ -1,6 +1,6 @@
 {
   "_class_name": "SanaTransformer2DModel",
-  "_diffusers_version": "0.32.0.dev0",
+  "_diffusers_version": "0.33.0.dev0",
   "attention_bias": false,
   "attention_head_dim": 32,
   "caption_channels": 1024,
@@ -8,6 +8,7 @@
   "cross_attention_head_dim": 112,
   "dropout": 0.0,
   "in_channels": 16,
+  "interpolation_scale": null,
   "mlp_ratio": 2.5,
   "norm_elementwise_affine": false,
   "norm_eps": 1e-06,

transformer/diffusion_pytorch_model.fp16.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b53b891f5e83b8f18a749b2e0a54d2c15b4fb7ee8acb956e6c8a9aae7b6b73a9
+oid sha256:36e6ecd58910031f5a3213f821e83a160e3ebf9c30eda15d6900660edf318339
 size 3203093344