Upload 4 files

LICENSE

@@ -0,0 +1,10 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+1. Definitions...
+
+(Full license text is truncated here, please include the full version from:
+https://www.apache.org/licenses/LICENSE-2.0.txt)

README.md

@@ -1,14 +1,44 @@
+# GNX Flux LoRA Inference (ComfyUI via Gradio)
+
+This Hugging Face Space runs a Flux .1 Dev LoRA model trained by @gauravjuneja using the ComfyUI framework.
+
+### ✨ Features
+- Custom prompt input
+- Runs locally with preloaded LoRA and UNet model
+- Inference pipeline powered by ComfyUI
+
+### 🔧 Usage
+1. Type a prompt like:
+   ```
+   gjnx is driving a Ferrari on a road in Germany
+   ```
+2. Wait for the CPU-based inference to return a generated image.
+
 ---
-title: GjnxGarry2
-emoji: 🖼
-colorFrom: purple
-colorTo: red
-sdk: gradio
-sdk_version: 5.25.2
-app_file: app.py
-pinned: false
-license: apache-2.0
-short_description: Lora
+
+### 📁 Model Structure
+
+This Space uses the following models:
+- `flux1-dev.sft` — UNet
+- `clip_l.safetensors` — CLIP-1
+- `t5xxl_fp16.safetensors` — CLIP-2
+- `ae.safetensors` — VAE
+- `gauravjuneja4.safetensors` — Trained LoRA
+
+Please ensure these are under the proper usage rights.
+
+---
+
+### 📜 License
+
+Code: [Apache 2.0](LICENSE)  
+LoRA Weights: Provided for educational/research use  
+Base Models: Original authors’ licenses apply (e.g., CreativeML Open RAIL-M if SD-derived)
+
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+### 🚀 Deployed With
+
+- [ComfyUI](https://github.com/comfyanonymous/ComfyUI)
+- [Gradio](https://www.gradio.app/)
+- [Hugging Face Spaces](https://huggingface.co/spaces)

app.py

@@ -1,154 +1,93 @@
-import gradio as gr
-import numpy as np
-import random
 
-# import spaces #[uncomment to use ZeroGPU]
-from diffusers import DiffusionPipeline
+import gradio as gr
+from gnx_flux_lora import import_custom_nodes, NODE_CLASS_MAPPINGS, get_value_at_index
 import torch
+import random
 
-device = "cuda" if torch.cuda.is_available() else "cpu"
-model_repo_id = "stabilityai/sdxl-turbo"  # Replace to the model you would like to use
-
-if torch.cuda.is_available():
-    torch_dtype = torch.float16
-else:
-    torch_dtype = torch.float32
-
-pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype)
-pipe = pipe.to(device)
-
-MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 1024
-
-
-# @spaces.GPU #[uncomment to use ZeroGPU]
-def infer(
-    prompt,
-    negative_prompt,
-    seed,
-    randomize_seed,
-    width,
-    height,
-    guidance_scale,
-    num_inference_steps,
-    progress=gr.Progress(track_tqdm=True),
-):
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-
-    generator = torch.Generator().manual_seed(seed)
-
-    image = pipe(
-        prompt=prompt,
-        negative_prompt=negative_prompt,
-        guidance_scale=guidance_scale,
-        num_inference_steps=num_inference_steps,
-        width=width,
-        height=height,
-        generator=generator,
-    ).images[0]
-
-    return image, seed
-
-
-examples = [
-    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
-    "An astronaut riding a green horse",
-    "A delicious ceviche cheesecake slice",
-]
-
-css = """
-#col-container {
-    margin: 0 auto;
-    max-width: 640px;
-}
-"""
-
-with gr.Blocks(css=css) as demo:
-    with gr.Column(elem_id="col-container"):
-        gr.Markdown(" # Text-to-Image Gradio Template")
-
-        with gr.Row():
-            prompt = gr.Text(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Enter your prompt",
-                container=False,
-            )
-
-            run_button = gr.Button("Run", scale=0, variant="primary")
-
-        result = gr.Image(label="Result", show_label=False)
-
-        with gr.Accordion("Advanced Settings", open=False):
-            negative_prompt = gr.Text(
-                label="Negative prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt",
-                visible=False,
-            )
-
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
-
-            with gr.Row():
-                width = gr.Slider(
-                    label="Width",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,  # Replace with defaults that work for your model
-                )
-
-                height = gr.Slider(
-                    label="Height",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=1024,  # Replace with defaults that work for your model
-                )
-
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.1,
-                    value=0.0,  # Replace with defaults that work for your model
-                )
-
-                num_inference_steps = gr.Slider(
-                    label="Number of inference steps",
-                    minimum=1,
-                    maximum=50,
-                    step=1,
-                    value=2,  # Replace with defaults that work for your model
-                )
-
-        gr.Examples(examples=examples, inputs=[prompt])
-    gr.on(
-        triggers=[run_button.click, prompt.submit],
-        fn=infer,
-        inputs=[
-            prompt,
-            negative_prompt,
-            seed,
-            randomize_seed,
-            width,
-            height,
-            guidance_scale,
-            num_inference_steps,
-        ],
-        outputs=[result, seed],
-    )
+# Initialize once (model loading outside inference function for Hugging Face ZeroGPU)
+import_custom_nodes()
+
+unetloader = NODE_CLASS_MAPPINGS["UNETLoader"]()
+unetloader_3 = unetloader.load_unet(
+    unet_name="flux1-dev.sft", weight_dtype="default"
+)
+
+dualcliploader = NODE_CLASS_MAPPINGS["DualCLIPLoader"]()
+dualcliploader_8 = dualcliploader.load_clip(
+    clip_name1="clip_l.safetensors",
+    clip_name2="t5xxl_fp16.safetensors",
+    type="flux",
+    device="default",
+)
+
+loraloader = NODE_CLASS_MAPPINGS["LoraLoader"]()
+loraloader_12 = loraloader.load_lora(
+    lora_name="gauravjuneja4/gauravjuneja4.safetensors",
+    strength_model=1,
+    strength_clip=1,
+    model=get_value_at_index(unetloader_3, 0),
+    clip=get_value_at_index(dualcliploader_8, 0),
+)
+
+vaeloader = NODE_CLASS_MAPPINGS["VAELoader"]()
+vae_9 = vaeloader.load_vae(vae_name="ae.safetensors")
+
+cliptextencode = NODE_CLASS_MAPPINGS["CLIPTextEncode"]()
+fluxguidance = NODE_CLASS_MAPPINGS["FluxGuidance"]()
+ksampler = NODE_CLASS_MAPPINGS["KSampler"]()
+vaedecode = NODE_CLASS_MAPPINGS["VAEDecode"]()
+emptylatentimage = NODE_CLASS_MAPPINGS["EmptyLatentImage"]()
+
+def run_inference(prompt: str):
+    with torch.inference_mode():
+        # Text encodings
+        positive = cliptextencode.encode(
+            text=prompt,
+            clip=get_value_at_index(loraloader_12, 1),
+        )
+
+        negative = cliptextencode.encode(
+            text="",
+            clip=get_value_at_index(loraloader_12, 1),
+        )
+
+        latent = emptylatentimage.generate(
+            width=1024, height=1024, batch_size=1
+        )
+
+        guided = fluxguidance.append(
+            guidance=3.5, conditioning=get_value_at_index(positive, 0)
+        )
+
+        sample = ksampler.sample(
+            seed=random.randint(1, 2**64),
+            steps=20,
+            cfg=1,
+            sampler_name="euler",
+            scheduler="simple",
+            denoise=1,
+            model=get_value_at_index(loraloader_12, 0),
+            positive=get_value_at_index(guided, 0),
+            negative=get_value_at_index(negative, 0),
+            latent_image=get_value_at_index(latent, 0),
+        )
+
+        decoded = vaedecode.decode(
+            samples=get_value_at_index(sample, 0),
+            vae=get_value_at_index(vae_9, 0),
+        )
+
+        result = get_value_at_index(decoded, 0)
+        return result
+
+# Gradio UI
+demo = gr.Interface(
+    fn=run_inference,
+    inputs=gr.Textbox(label="Prompt", placeholder="e.g. gjnx is driving ferrari on a road in germany"),
+    outputs=gr.Image(label="Generated Image"),
+    title="GNX Flux LoRA Generator",
+    description="Enter a prompt using your trained LoRA with Flux .1 Dev."
+)
 
 if __name__ == "__main__":
     demo.launch()

requirements.txt

@@ -1,6 +1,5 @@
-accelerate
-diffusers
-invisible_watermark
 torch
+gradio
+comfyui
 transformers
-xformers
+diffusers