Add suggested inference code

README.md

@@ -97,5 +97,102 @@ Krea realtime allows users to generate videos in a streaming fashion with ~1s ti
     </table>
 </div>
 
+# Use it with our inference code
 
+Set up
+```bash
+sudo apt install ffmpeg # install if you haven't already
+git clone https://github.com/krea-ai/realtime-video
+cd realtime-video
+uv sync
+uv pip install flash_attn --no-build-isolation
+huggingface-cli download Wan-AI/Wan2.1-T2V-1.3B --local-dir-use-symlinks False --local-dir wan_models/Wan2.1-T2V-1.3B
+huggingface-cli download krea/krea-realtime-video krea-realtime-video-14b.safetensors --local-dir-use-symlinks False --local-dir checkpoints/krea-realtime-video-14b.safetensors
+```
 
+Run
+```bash
+export MODEL_FOLDER=Wan-AI
+export CUDA_VISIBLE_DEVICES=0 # pick the GPU you want to serve on
+export DO_COMPILE=true
+
+uvicorn release_server:app --host 0.0.0.0 --port 8000
+``` 
+
+And use the web app at http://localhost:8000/ in your browser 
+(for more advanced use-cases and custom pipeline check out our GitHub repository: https://github.com/krea-ai/realtime-video)
+
+# Use it with 🧨 diffusers
+
+Krea Realtime 14B can be used with the `diffusers` library utilizing the new Modular Diffusers structure (for now supporting text-to-video, video-to-video coming soon)
+
+```bash
+# Install diffusers from main
+pip install git+github.com/huggingface/diffusers.git
+``` 
+
+```py
+import torch
+from collections import deque
+from diffusers import ModularPipelineBlocks, FlowMatchEulerDiscreteScheduler
+from diffusers.utils import export_to_video
+from diffusers.modular_pipelines import PipelineState, WanModularPipeline
+
+class WanRTStreamingPipeline(WanModularPipeline):
+    @property
+    def default_sample_height(self):
+        return 60
+
+    @property
+    def default_sample_width(self):
+        return 104
+
+    @property
+    def frame_seq_length(self):
+        return 1560
+
+    @property
+    def seq_length(self):
+        return 32760
+
+    @property
+    def kv_cache_num_frames(self):
+        return 3
+
+    @property
+    def frame_cache_len(self):
+        return 1 + (self.kv_cache_num_frames - 1) * 4
+
+
+block_path = "krea/krea-realtime-video"
+blocks = ModularPipelineBlocks.from_pretrained(block_path, trust_remote_code=True)
+pipe = WanRTStreamingPipeline(blocks, block_path)
+
+pipe.load_components(
+    trust_remote_code=True,
+    device_map="cuda",
+    torch_dtype={"default": torch.bfloat16, "vae": torch.float32},
+)
+pipe.scheduler = FlowMatchEulerDiscreteScheduler(shift=5.0)
+
+prompt = ["A cat sitting on a boat"]
+
+num_frames_per_block = 3
+num_blocks = 9
+
+frames = []
+state = PipelineState()
+state.set("frame_cache_context", deque(maxlen=pipe.frame_cache_len))
+for block_idx in range(num_blocks):
+    state = pipe(
+        state,
+        prompt=prompt,
+        num_inference_steps=6,
+        num_blocks=num_blocks,
+        num_frames_per_block=num_frames_per_block,
+        block_idx=block_idx,
+    )
+    frames.extend(state.values["videos"][0])
+
+export_to_video(frames, "krt.mp4")
+```