Initial commit with folder contents

pyproject.toml

@@ -0,0 +1,27 @@
+[build-system]
+requires = ["setuptools >= 75.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "flux-schnell-edge-inference"
+description = "An edge-maxxing model submission for the 4090 Flux contest"
+requires-python = ">=3.10,<3.13"
+version = "7"
+dependencies = [
+  "diffusers==0.31.0",
+  "transformers==4.46.2",
+  "accelerate==1.1.0",
+  "omegaconf==2.3.0",
+  "torch==2.5.1",
+  "protobuf==5.28.3",
+  "sentencepiece==0.2.0",
+  "edge-maxxing-pipelines @ git+https://github.com/womboai/edge-maxxing@7c760ac54f6052803dadb3ade8ebfc9679a94589#subdirectory=pipelines",
+  "gitpython>=3.1.43",
+  "torchao>=0.6.1"
+]
+
+[tool.edge-maxxing]
+models = ["black-forest-labs/FLUX.1-schnell"]
+
+[project.scripts]
+start_inference = "main:main"

src/main.py

@@ -0,0 +1,59 @@
+import atexit
+from io import BytesIO
+from multiprocessing.connection import Listener
+from os import chmod, remove
+from os.path import abspath, exists
+from pathlib import Path
+
+import torch
+
+from PIL.JpegImagePlugin import JpegImageFile
+from pipelines.models import TextToImageRequest
+
+from pipeline import load_pipeline, infer
+
+SOCKET = abspath(Path(__file__).parent.parent / "inferences.sock")
+
+
+def at_exit():
+    torch.cuda.empty_cache()
+
+
+def main():
+    atexit.register(at_exit)
+
+    print(f"Loading pipeline")
+    pipeline = load_pipeline()
+
+    print(f"Pipeline loaded! , creating socket at '{SOCKET}'")
+
+    if exists(SOCKET):
+        remove(SOCKET)
+
+    with Listener(SOCKET) as listener:
+        chmod(SOCKET, 0o777)
+
+        print(f"Awaiting connections")
+        with listener.accept() as connection:
+            print(f"Connected")
+
+            while True:
+                try:
+                    request = TextToImageRequest.model_validate_json(connection.recv_bytes().decode("utf-8"))
+                except EOFError:
+                    print(f"Inference socket exiting")
+
+                    return
+
+                image = infer(request, pipeline)
+
+                data = BytesIO()
+                image.save(data, format=JpegImageFile.format)
+
+                packet = data.getvalue()
+
+                connection.send_bytes(packet)
+
+
+if __name__ == '__main__':
+    main()

src/pipeline.py

@@ -0,0 +1,85 @@
+from diffusers import FluxPipeline, AutoencoderKL, AutoencoderTiny
+from diffusers.schedulers import FlowMatchEulerDiscreteScheduler
+
+from transformers import T5EncoderModel, T5TokenizerFast, CLIPTokenizer, CLIPTextModel
+import torch
+import torch._dynamo
+import gc
+from PIL import Image as img
+from PIL import Image
+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 diffusers.image_processor import VaeImageProcessor
+Pipeline = None
+import os
+MODEL_ID = "black-forest-labs/FLUX.1-schnell"
+traced_vae_decode_path = "traced_vae_decode.pt"  
+def empty_cache():
+    start = time.time()
+    gc.collect()
+    torch.cuda.empty_cache()
+    torch.cuda.reset_max_memory_allocated()
+    torch.cuda.reset_peak_memory_stats()
+    print(f"Flush took: {time.time() - start}")
+
+
+
+def load_pipeline() -> Pipeline:    
+    empty_cache()
+    dtype, device = torch.bfloat16, "cuda"
+    vae = AutoencoderKL.from_pretrained(
+        MODEL_ID, subfolder="vae", torch_dtype=torch.bfloat16
+    )
+    quantize_(vae, int8_weight_only())
+    pipeline = DiffusionPipeline.from_pretrained(
+        MODEL_ID,
+        vae=vae,
+        torch_dtype=dtype,
+        )
+        
+    pipeline.enable_sequential_cpu_offload()
+    for _ in range(2):
+        empty_cache()
+        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)
+    
+    return pipeline
+
+def trace_and_save_vae_decoder(vae, latents):
+    try:
+        traced_vae_decode = torch.jit.trace(vae.decode, (latents, True))
+        torch.jit.save(traced_vae_decode, traced_vae_decode_path)
+        return traced_vae_decode
+    except Exception as e:
+        print(f"JIT tracing failed: {e}")
+        return vae.decode #Fall back to untraced decoder.
+        
+def decode_latents_to_image(latents, height: int, width: int, vae):
+    if not height:
+        height = 1024
+    if not width:
+        width = 1024
+    if vae.config.block_out_channels:
+        vae_scale_factor = 2 ** (len(vae.config.block_out_channels))
+    else:
+        vae_scale_factor = 1
+    image_processor = VaeImageProcessor(vae_scale_factor=vae_scale_factor)
+
+    traced_vae_decode = vae.decode
+    with torch.no_grad():
+        latents = FluxPipeline._unpack_latents(latents.unsqueeze(0), height, width, vae_scale_factor)
+        latents = (latents / vae.config.scaling_factor) + vae.config.shift_factor
+        image = traced_vae_decode(latents, return_dict=False)[0]  # Use the traced function
+        decoded_image = image_processor.postprocess(image, output_type="pil")[0]
+        
+    return decoded_image
+
+
+@torch.inference_mode()
+def infer(request: TextToImageRequest, pipeline: Pipeline) -> Image:
+    empty_cache()
+    generator = Generator("cuda").manual_seed(request.seed)
+    latent=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="latent").images[0]
+    return decode_latents_to_image(latent, request.height, request.width, pipeline.vae)