new update fix

README.md

@@ -11,8 +11,239 @@ pinned: false
 license: apache-2.0
 short_description: Mincraftify converts all images into mc-style LIVE!
 ---
+# Minecraftify!
 
-# ⛏️ Minecraft Spatial Voxel Filter
-Real-time video stream processing pipeline using FastRTC and Modal serverless environments.
+**Minecraftify your images live.**
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+Minecraftify is a Hugging Face Gradio Space that turns uploaded photos into a faithful vanilla Minecraft interpretation of the same scene. It is powered by a fine-tuned **FLUX.2-Klein-4B** img2img LoRA trained on a custom dataset generated with **Qwen-Edit-25-12**.
+
+## Live Demo
+
+* **Space:** Minecraftify
+* **Demo video:** [YouTube walkthrough](#)
+* **Blog post:** [Project blog](#)
+
+## Current Status
+
+Playable demo Space with:
+
+* still-image Minecraftification
+* live webcam mode
+* persistent model caching on Space storage
+* LoRA-based FLUX.2-Klein inference
+* Gradio UI with image upload, webcam input, and advanced settings
+
+## What It Does
+
+Minecraftify transforms an input image into a Minecraft-style version of the same scene while trying to preserve:
+
+* composition
+* camera angle
+* layout
+* objects already present in the scene
+* color relationships and overall structure
+
+The model is tuned to:
+
+* replace realistic surfaces with Minecraft blocks and voxel geometry
+* keep the scene recognizable
+* avoid unnecessary scene changes
+* convert people, animals, and objects into Minecraft-style equivalents where needed
+
+## Project Artifacts
+
+* **Base model:** `black-forest-labs/FLUX.2-klein-4B`
+* **LoRA adapter:** `AnimeOverlord/flux2-klein-4b-mc-v2`
+* **Dataset:** 400 image pairs created with Qwen-Edit-25-12
+* **Training script:** `train_dreambooth_lora_flux2_klein_img2img.py`
+
+## Fine-Tuning Setup
+
+The LoRA was trained with FLUX.2-Klein img2img using a paired dataset with:
+
+* `source_image` as the conditioning image
+* `edited_image` as the target image
+* `prompt_used` as the caption column
+
+Training highlights:
+
+* `train_batch_size=1`
+* `gradient_accumulation_steps=4`
+* `mixed_precision=bf16`
+* `learning_rate=2e-6`
+* `lr_scheduler=constant_with_warmup`
+* `lr_warmup_steps=200`
+* `max_train_steps=1200`
+* `rank=64`
+* `cache_latents`
+* `use_8bit_adam`
+* `aspect_ratio_buckets` enabled
+
+## Hackathon Fit
+
+Minecraftify is designed as a compact, fun, image-to-image Space with a strong visual identity and an immediate demo loop.
+
+It fits the small-model spirit because the core generation path is built around a **4B FLUX Klein model** with a LoRA adapter rather than a large general-purpose model.
+
+## How to Demo
+
+1. Upload an image or start the webcam.
+2. Choose still image or live mode.
+3. Adjust inference steps, guidance scale, and seed.
+4. Click **Minecraftify!**
+5. Download or inspect the result.
+
+## Recommended Demo Settings
+
+* **Inference steps:** 3
+* **Guidance scale:** 3.0
+* **Seed:** any fixed value for reproducibility
+* **Input:** well-lit images with clear objects and simple scenes
+
+## Features
+
+* image upload
+* webcam capture
+* live frame processing
+* prompt-based scene preservation
+* persistent model caching in Hugging Face Space storage
+* adjustable inference settings
+* output comparison view
+
+## Model and Runtime
+
+The app loads the FLUX.2-Klein base model and then applies the Minecraft LoRA adapter.
+
+Runtime behavior:
+
+* models are cached on persistent Space storage
+* weights are reused across runs when present
+* the pipeline is kept in memory for the active session
+* image generation uses img2img inference for scene preservation
+
+## Space Storage
+
+This Space is configured to use persistent storage so model files do not need to be downloaded every time the Space restarts.
+
+* model cache path: `/data/models`
+* Hugging Face cache path: `/data/.huggingface`
+
+## Architecture
+
+**Input image** → **FLUX.2-Klein img2img** → **Minecraft LoRA** → **Rendered output**
+
+For live mode, webcam frames are captured continuously and only the latest frame is processed when the model becomes available.
+
+## Local Development
+
+This project was trained locally with PyTorch and Accelerate, and the training workflow also supports pushing the fine-tuned model to the Hugging Face Hub.
+
+### Running locally with PyTorch
+
+#### 1) Install the training dependencies
+
+For the most up-to-date Diffusers example scripts, it is recommended to install Diffusers from source:
+
+```bash
+git clone https://github.com/huggingface/diffusers
+cd diffusers
+pip install -e .
+````
+
+Then install the FLUX DreamBooth example requirements:
+
+```bash
+cd examples/dreambooth
+pip install -r requirements_flux.txt
+```
+
+#### 2) Configure Accelerate
+
+Set up Accelerate for your environment:
+
+```bash
+accelerate config
+```
+
+If you want the default configuration without answering prompts:
+
+```bash
+accelerate config default
+```
+
+If you are running in a notebook or another environment without an interactive shell:
+
+```python
+from accelerate.utils import write_basic_config
+write_basic_config()
+```
+
+If possible, enable `torch compile` in Accelerate for faster training. Also make sure `peft>=0.6.0` is installed, since PEFT is used as the LoRA backend.
+
+### Training FLUX.2-Klein LoRA on an image-to-image dataset
+
+This project uses the FLUX 2 Klein 4B base model and trains a LoRA adapter on a paired img2img dataset.
+
+```bash
+cd diffusers/examples/dreambooth && accelerate launch train_dreambooth_lora_flux2_klein_img2img.py \
+  --pretrained_model_name_or_path=black-forest-labs/FLUX.2-klein-4B \
+  --output_dir="flux2-i2i" \
+  --dataset_name="AnimeOverlord/mine-dataset" \
+  --image_column="edited_image" \
+  --cond_image_column="source_image" \
+  --caption_column="prompt_used" \
+  --gradient_checkpointing \
+  --cache_latents \
+  --train_batch_size=1 \
+  --guidance_scale=1 \
+  --gradient_accumulation_steps=4 \
+  --mixed_precision="bf16" \
+  --optimizer="prodigy" \
+  --learning_rate=1 \
+  --lr_warmup_steps=200 \
+  --max_train_steps=1200 \
+  --rank=64 \
+  --seed="0" \
+  --push_to_hub \
+  --hub_model_id="[YOURACCOUNT]/flux2-klein-4b-mc" \
+  --aspect_ratio_buckets="672,1568;688,1504;720,1456;752,1392;800,1328;832,1248;880,1184;944,1104;1024,1024;1104,944;1184,880;1248,832;1328,800;1392,752;1456,720;1504,688;1568,672"
+```
+
+### Notes
+
+* The dataset contains **400 images** created with **Qwen-Edit-25-12**.
+* The training run uses a paired img2img setup with:
+
+  * `source_image` as the conditioning image
+  * `edited_image` as the target image
+  * `prompt_used` as the caption
+* `push_to_hub` uploads the trained LoRA adapter to the Hugging Face Hub.
+* The aspect-ratio buckets help keep training efficient across different image shapes.
+
+### Output
+
+After training, the LoRA adapter is published to:
+
+```bash
+YOURACCOUNT/flux2-klein-4b-mc
+```
+
+## Blog and Video Links
+
+* **Blog:** [Read the build notes](#)
+* **YouTube:** [Watch the walkthrough](#)
+
+## Link to Notebooks Used
+* **Training Notbook:** [Modal Notebook](https://modal.com/notebooks/kalamkhan-se/main/nb-ygLQVGDvJR3FbrNpQwGYfV)
+* **Data Creation Notebook:** [Modal Notebook](https://modal.com/notebooks/kalamkhan-se/main/nb-L22FAz1tYTB39h7tXkXF4N)
+
+## Credits
+
+* Base model: Black Forest Labs
+* Fine-tuning workflow: Hugging Face Diffusers
+* Dataset creation: Qwen-Edit-25-12
+* UI: Gradio
+
+## License
+
+This project is a demo Space for experimentation and presentation. Check the model and dataset licenses before redistribution.

app.py

@@ -643,7 +643,7 @@ input[type="range"]::-webkit-slider-thumb {
 # ----------------------------------------------------------------------
 
 BASE_MODEL_ID = "black-forest-labs/FLUX.2-klein-4B"
-LORA_ID = "AnimeOverlord/flux2-klein-4b-mc"
+LORA_ID = "AnimeOverlord/flux2-klein-4b-mc-v2"
 
 # Mount this path in the Space settings
 SPACE_DATA_DIR = "/data"
@@ -1000,14 +1000,11 @@ with gr.Blocks(**BLOCKS_KWARGS) as demo:
 
                 with gr.Accordion("⚙️ Advanced Settings", open=False):
                     steps = gr.Slider(
-                        minimum=2, maximum=15, value=3, step=1,
+                        minimum=1, maximum=15, value=2, step=1,
                         label="⚡ Inference Steps",
                         info="Higher = better quality, slower generation",
                     )
-                    guidance_scale = gr.Slider(
-                        minimum=1.0, maximum=10.0, value=3.0, step=0.5,
-                        label="🧭 Guidance Scale",
-                    )
+
                     seed = gr.Number(
                         value=42, precision=0,
                         label="🎲 Seed",
@@ -1029,12 +1026,12 @@ with gr.Blocks(**BLOCKS_KWARGS) as demo:
 
         # ── RIGHT: Output Column ──
         with gr.Column(scale=5):
-            gr.HTML('<div class="mc-section-label"><span class="mc-icon">🧱</span> MINECRAFT VERSION</div>')
+            gr.HTML('<div class="mc-section-label"><span class="mc-icon">🧱</span> MINECRAFT STYLE</div>')
 
             with gr.Group(elem_classes="mc-panel"):
                 output_image = gr.Image(
                     type="pil",
-                    label="Minecraft version",
+                    label="Minecraft Style",
                     interactive=False,
                     height=400,
                 )
@@ -1045,7 +1042,7 @@ with gr.Blocks(**BLOCKS_KWARGS) as demo:
             <div class="mc-tip-title">📖 TIPS & TRICKS</div>
             <div class="mc-divider" style="margin: 10px 0;"></div>
             <ul style="padding-left: 20px; margin: 8px 0;">
-                <li>🎯 <strong style="color:#7CB342;">Inference steps:</strong> 3 is the sweet spot. Go higher for more detail.</li>
+                <li>🎯 <strong style="color:#7CB342;">Inference steps:</strong> 2 is the sweet spot. Go higher for more detail and Craziness!</li>
                 <li>🐺 <strong style="color:#7CB342;">Still too real?</strong> Describe tricky parts in the details box (e.g. "convert the dog into a wolf mob").</li>
                 <li>⏱️ <strong style="color:#7CB342;">Live mode:</strong> The camera stream prioritizes the newest frame whenever the model is free.</li>
                 <li>🎲 <strong style="color:#7CB342;">Variations:</strong> Change the seed to get different Minecraft interpretations of the same photo.</li>
@@ -1079,7 +1076,7 @@ with gr.Blocks(**BLOCKS_KWARGS) as demo:
 
     run_btn.click(
         fn=minecraftify,
-        inputs=[input_image, steps, guidance_scale, seed, extra_details],
+        inputs=[input_image, steps, 1.0, seed, extra_details],
         outputs=output_image,
     )
 
@@ -1092,7 +1089,7 @@ with gr.Blocks(**BLOCKS_KWARGS) as demo:
         queue=False,
     ).then(
         fn=minecraftify_live_loop,
-        inputs=[steps, guidance_scale, seed, extra_details],
+        inputs=[steps, 1.0, seed, extra_details],
         outputs=[output_image],
         concurrency_limit=1,
     )