Update src/pipeline.py

src/pipeline.py

@@ -1,8 +1,5 @@
-from diffusers import FluxPipeline, AutoencoderKL, AutoencoderTiny
-from diffusers.image_processor import VaeImageProcessor
-from diffusers.schedulers import FlowMatchEulerDiscreteScheduler
-
-from transformers import T5EncoderModel, T5TokenizerFast, CLIPTokenizer, CLIPTextModel
+from diffusers import AutoencoderTiny
+from transformers import T5EncoderModel
 import torch
 import torch._dynamo
 import gc
@@ -12,7 +9,7 @@ from pipelines.models import TextToImageRequest
 from torch import Generator
 import time
 from diffusers import FluxTransformer2DModel, DiffusionPipeline
-from torchao.quantization import quantize_,int8_weight_only
+from torchao.quantization import quantize_, int8_weight_only
 import os
 os.environ["PYTORCH_CUDA_ALLOC_CONF"] = "expandable_segments:False,garbage_collection_threshold:0.01"
 
@@ -29,35 +26,40 @@ def load_pipeline() -> Pipeline:
     clear()
    
     text_encoder_2 = T5EncoderModel.from_pretrained(
-        "city96/t5-v1_1-xxl-encoder-bf16", torch_dtype=torch.bfloat16
+        "city96/t5-v1_1-xxl-encoder-bf16", torch_dtype=DTYPE
     )
-    vae=AutoencoderKL.from_pretrained(ckpt_id, subfolder="vae", torch_dtype=dtype)
+    
+    vae = AutoencoderTiny.from_pretrained("madebyollin/taef1", torch_dtype=DTYPE)
+    quantize_(vae, int8_weight_only())
     pipeline = DiffusionPipeline.from_pretrained(
-        ckpt_id,
+        MODEL_ID,
         vae=vae,
         text_encoder_2 = text_encoder_2,
-        torch_dtype=dtype,
+        torch_dtype=DTYPE,
         )
     torch.backends.cudnn.benchmark = True
     torch.backends.cuda.matmul.allow_tf32 = True
     torch.cuda.set_per_process_memory_fraction(0.99)
     pipeline.text_encoder.to(memory_format=torch.channels_last)
     pipeline.transformer.to(memory_format=torch.channels_last)
-    
-
     pipeline.vae.to(memory_format=torch.channels_last)
     pipeline.vae = torch.compile(pipeline.vae)
     
     pipeline._exclude_from_cpu_offload = ["vae"]
     pipeline.enable_sequential_cpu_offload()
-    for _ in range(2):
+    for _ in range(1):
         pipeline(prompt="onomancy, aftergo, spirantic, Platyhelmia, modificator, drupaceous, jobbernowl, hereness", width=1024, height=1024, guidance_scale=0.0, num_inference_steps=4, max_sequence_length=256)
-    
+    pipeline(prompt="", width=1024, height=1024, guidance_scale=0.0, num_inference_steps=4, max_sequence_length=256)
     return pipeline
 
 
+sample = True
 @torch.inference_mode()
 def infer(request: TextToImageRequest, pipeline: Pipeline) -> Image:
+    global sample
+    if sample:
+        clear()
+        sample = None
     torch.cuda.reset_peak_memory_stats()
     generator = Generator("cuda").manual_seed(request.seed)
     image=pipeline(request.prompt,generator=generator, guidance_scale=0.0, num_inference_steps=4, max_sequence_length=256, height=request.height, width=request.width, output_type="pil").images[0]