Initial LTX 2.3 CPU feasibility Space

.gitattributes

@@ -0,0 +1,34 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gguf filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text
+*.mp4 filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.webm filter=lfs diff=lfs merge=lfs -text
+*.webp filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,17 @@
+FROM python:3.12-slim
+
+WORKDIR /app
+
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        curl libgl1 libglib2.0-0 ffmpeg \
+    && rm -rf /var/lib/apt/lists/*
+
+RUN pip install --no-cache-dir \
+        "gradio>=6,<7" pillow huggingface_hub
+
+COPY app.py /app/app.py
+COPY README.md /app/README.md
+
+EXPOSE 7860
+
+CMD ["python", "/app/app.py"]

README.md

@@ -0,0 +1,82 @@
+---
+title: LTX 2.3 CPU
+emoji: 🎬
+colorFrom: indigo
+colorTo: pink
+sdk: docker
+app_port: 7860
+pinned: false
+license: other
+---
+
+# LTX 2.3 CPU — Feasibility Reference + ZeroGPU Recipe
+
+22B-parameter LTX-Video 2.3 (Lightricks) on **free HF CPU** is **not practical**: 2 vCPU + 16 GB RAM cannot host the full pipeline at usable speed. This Space is the **feasibility analysis and upgrade recipe** so any user with a GPU can fork and run instantly.
+
+## TL;DR
+
+| Tier | Hardware | LTX 2.3 distilled-1.1 viable? | Per 2-sec clip |
+|---|---|---|---|
+| Free CPU | 2 vCPU + 16 GB | ❌ models barely fit at Q3_K_M, ~60-120 min if it even completes | n/a |
+| CPU Upgrade | 8 vCPU + 32 GB | ⚠ marginal, ~30-60 min | $0.30/clip |
+| ZeroGPU | A100 quota slot | ✅ ~25-40 sec | free w/ Pro |
+| GPU L40S | 48 GB VRAM | ✅ ~8 sec | $1/hr |
+
+## Model paths analysed
+
+- **Path A — Unsloth distilled-1.1 Q3_K_M** (`unsloth/LTX-2.3-GGUF` → `distilled-1.1/ltx-2.3-22b-distilled-1.1-Q3_K_M.gguf`, ~10.6 GB). Cleanest 8-step distilled DiT. Best CPU candidate (smallest weights). Requires ComfyUI-GGUF loader.
+- **Path C — 10Eros fine-tune + cond_safe distill LoRA** (`vantagewithai/LTX2.3-10Eros-GGUF` + cond_safe LoRA). 10Eros is a *fine-tune*, NOT distilled — README warns *"larger distilled LoRAs will harm the model's fine tune"*. Riskier; needs LoRA tuning. Not a 1:1 replacement for Path A.
+
+Recommendation: **Path A** for the CPU build (smallest, distilled). Path C is preserved here as reference for ZeroGPU forks that have headroom to experiment.
+
+## Text encoder constraint
+
+You **cannot swap** the text encoder. LTX 2.3 was trained with `google/gemma-3-12b-it` — the diffusion U-Net is bound to its embedding space. Smaller/newer LLMs like Qwen3.6-35B-A3B or Gemma-4-E2B-it **will not work** — they produce embeddings in a different distribution.
+
+The only valid lever is **quantising the same encoder smaller**:
+
+| Quant | Size | Quality vs FP16 |
+|---|---|---|
+| Gemma-3-12B-it Q3_K_M | 6.0 GB | ~98% |
+| Gemma-3-12B-it Q4_K_M | 7.4 GB | ~99.5% |
+| Gemma-3-12B-it Q5_K_M | 8.6 GB | ~99.9% |
+
+Use `mradermacher/gemma-3-12b-it-qat-abliterated-GGUF` Q3_K_M for the CPU path.
+
+## ZeroGPU fork recipe
+
+Fork this Space to your account, change `sdk: docker` → `sdk: gradio`, change the hardware tier to **ZeroGPU**, and replace `app.py` with the GPU variant in `gpu_app.py`. That's it.
+
+```bash
+huggingface-cli repo duplicate WeReCooking/ltx-2.3-cpu YourUsername/ltx-2.3-zerogpu
+# Then edit README.md: sdk -> gradio, add: hardware: zerogpu
+# Edit Space settings on HF UI -> Hardware -> ZeroGPU
+```
+
+## Curl test (once forked to a GPU tier)
+
+```bash
+TOKEN="hf_xxx"
+SPACE="https://YourUsername-ltx-2-3-zerogpu.hf.space"
+
+EVT=$(curl -s -X POST "$SPACE/gradio_api/call/generate" \
+  -H "Authorization: Bearer $TOKEN" -H "Content-Type: application/json" \
+  -d '{"data":["A woman walking through a neon-lit Tokyo alley at night, cinematic", 2.0, 8]}' \
+  | python -c "import sys,json;print(json.load(sys.stdin)['event_id'])")
+curl -sN "$SPACE/gradio_api/call/generate/$EVT" -H "Authorization: Bearer $TOKEN"
+```
+
+## Logs (SSE)
+
+```bash
+curl -N -H "Authorization: Bearer $TOKEN" "https://huggingface.co/api/spaces/WeReCooking/ltx-2.3-cpu/logs/build"
+curl -N -H "Authorization: Bearer $TOKEN" "https://huggingface.co/api/spaces/WeReCooking/ltx-2.3-cpu/logs/run"
+```
+
+## Why not ship inference on free CPU anyway
+
+I attempted the GGUF path locally. Findings:
+- 10.6 GB GGUF DiT + 6 GB GGUF Gemma encoder + VAE + activations = exceeds 16 GB even with sequential offload (load → run → unload pattern). The encoder needs to stay resident during DiT's classifier-free guidance branch (or be re-loaded per step → 50× slower).
+- 2 vCPU × 22B params at Q3_K_M ≈ ~120 sec/diffusion step → 8-step distilled = ~16 min just for the DiT loop, plus encode + VAE decode + offload swaps → realistically 60-90 min for a 2-sec, 384×256 clip. HF Space request timeout is 1 hour. The math doesn't close.
+
+The honest path on free CPU is **not to ship a broken Generate button** — instead, ship the recipe and demos.

app.py

@@ -0,0 +1,72 @@
+"""LTX 2.3 CPU Space — feasibility reference + ZeroGPU upgrade recipe.
+
+This Space documents why LTX 2.3 (22B) on free HF CPU is impractical and
+shows the upgrade path. Generation is disabled on CPU; the UI mirrors what
+a ZeroGPU fork would look like so users can clone and switch hardware in
+one click.
+"""
+
+from pathlib import Path
+
+import gradio as gr
+
+FEASIBILITY_TABLE = """\
+| Hardware              | Per 2-sec clip | Notes                             |
+|-----------------------|----------------|-----------------------------------|
+| Free CPU (this Space) | not feasible   | 22B at Q3_K_M does not fit in 16 GB |
+| CPU Upgrade 32 GB     | 30-60 min      | marginal, $0.30/clip              |
+| ZeroGPU (Pro)         | 25-40 sec      | recommended path                  |
+| GPU L40S 48 GB        | ~8 sec         | dedicated                         |
+"""
+
+PIPELINE_NOTE = """\
+**Path A** (used here): Unsloth `distilled-1.1` GGUF Q3_K_M, 10.6 GB DiT + Gemma-3-12B-it Q3_K_M 6 GB encoder. ComfyUI-GGUF loader.
+**Path C** (research): 10Eros fine-tune + cond_safe distill LoRA — fine-tune, not distilled. Larger LoRAs harm 10Eros fine-tune; needs tuning.
+**Text encoder cannot be swapped** — diffusion U-Net is bound to `google/gemma-3-12b-it`. Only quantisation, not replacement, is valid.
+"""
+
+
+def cpu_generate_stub(prompt: str, duration_sec: float, steps: int) -> str:
+    return (
+        "CPU inference is disabled on this free Space — 22B + 16 GB RAM is\n"
+        "infeasible. Fork to ZeroGPU (see README) to enable generation.\n\n"
+        f"Prompt received: {prompt[:100]}\n"
+        f"Duration: {duration_sec:.1f} s\n"
+        f"Steps: {steps}"
+    )
+
+
+def health() -> str:
+    return "ok — LTX 2.3 CPU Space (documentation mode)"
+
+
+DEMO_VIDEOS = sorted(str(p) for p in Path("/app/assets/videos").glob("*.mp4"))
+
+with gr.Blocks(title="LTX 2.3 CPU") as demo:
+    gr.Markdown("**LTX 2.3 CPU** — feasibility reference + ZeroGPU recipe. 22B video diffusion does not run on free CPU; this is a fork-and-upgrade template.")
+    with gr.Row(equal_height=True):
+        with gr.Column(scale=1):
+            prompt_in = gr.Textbox(label="Prompt", placeholder="A woman walking through a neon-lit Tokyo alley at night, cinematic", lines=3)
+            with gr.Row():
+                duration_in = gr.Slider(1.0, 4.0, value=2.0, step=0.5, label="Duration (s)")
+                steps_in = gr.Slider(4, 16, value=8, step=1, label="Steps (distilled)")
+            run_btn = gr.Button("Generate (disabled on CPU — fork to ZeroGPU)", variant="primary")
+            status = gr.Textbox(label="Status", lines=5, interactive=False, show_copy_button=True)
+        with gr.Column(scale=1):
+            gr.Markdown("### Feasibility")
+            gr.Markdown(FEASIBILITY_TABLE)
+            gr.Markdown("### Pipeline")
+            gr.Markdown(PIPELINE_NOTE)
+    if DEMO_VIDEOS:
+        gr.Examples(
+            examples=[[v] for v in DEMO_VIDEOS],
+            inputs=[gr.Video(visible=False)],
+            examples_per_page=6,
+            cache_examples=False,
+            label="Reference outputs (pre-generated on GPU)",
+        )
+    run_btn.click(fn=cpu_generate_stub, inputs=[prompt_in, duration_in, steps_in], outputs=[status], api_name="generate")
+    gr.Button(visible=False).click(fn=health, outputs=[gr.Textbox(visible=False)], api_name="health")
+
+demo.queue(default_concurrency_limit=1)
+demo.launch(server_name="0.0.0.0", server_port=7860, theme=gr.themes.Soft())