Authenticate the gated VAE download

app.py

@@ -21,16 +21,16 @@ import time
 from fractions import Fraction
 from pathlib import Path
 
+# Roughly 22 GB arrives on a cold start, and `huggingface_hub` reads this into
+# a module constant at import - setting it any later has no effect at all.
+os.environ.setdefault("HF_HUB_ENABLE_HF_TRANSFER", "1")
+
 import gradio as gr
 import spaces
 from huggingface_hub import hf_hub_download, snapshot_download
 
 # --------------------------------------------------------------------- layout
 
-# Roughly 22 GB arrives on a cold start; the accelerated transfer is the
-# difference between a few minutes and a startup timeout.
-os.environ.setdefault("HF_HUB_ENABLE_HF_TRANSFER", "1")
-
 ROOT = Path(__file__).resolve().parent
 WORK = Path(os.environ.get("LTX_WORK", "/tmp/ltx"))
 COMFY = WORK / "ComfyUI"
@@ -70,26 +70,37 @@ def _clone(url: str, dest: Path, commit: str) -> None:
     _run(["git", "checkout", commit], cwd=dest)
 
 
+#: `Lightricks/LTX-2.5` is gated, so the VAEs need an authenticated request from
+#: an account that has accepted the licence. Set HF_TOKEN as a Space secret.
+TOKEN = os.environ.get("HF_TOKEN") or None
+
+
 def bootstrap() -> dict[str, Path]:
     """Everything that does not need a GPU, done once at import."""
+    if TOKEN is None:
+        raise RuntimeError(
+            "HF_TOKEN is not set. Lightricks/LTX-2.5 is a gated repo and its "
+            "VAEs cannot be fetched without a token from an account that has "
+            "accepted the licence.")
     _clone("https://github.com/comfyanonymous/ComfyUI.git", COMFY, COMFY_COMMIT)
     _clone("https://github.com/city96/ComfyUI-GGUF.git",
            COMFY / "custom_nodes" / "ComfyUI-GGUF", GGUF_NODE_COMMIT)
     MODELS.mkdir(parents=True, exist_ok=True)
     OUT.mkdir(parents=True, exist_ok=True)
 
-    paths = {
-        "packed": Path(hf_hub_download(ENCODER_REPO, ENCODER_FILE,
-                                       cache_dir=str(MODELS))),
-        "encoder_dir": Path(snapshot_download(ENCODER_REPO, allow_patterns=["encoder-hf/*"],
-                                              cache_dir=str(MODELS))) / "encoder-hf",
-        "gguf": Path(hf_hub_download(DIT_REPO, DIT_FILE, cache_dir=str(MODELS))),
-        "video_vae": Path(hf_hub_download(VAE_REPO, VIDEO_VAE, revision=VAE_REVISION,
-                                          cache_dir=str(MODELS))),
-        "audio_vae": Path(hf_hub_download(VAE_REPO, AUDIO_VAE, revision=VAE_REVISION,
-                                          cache_dir=str(MODELS))),
+    def fetch(repo: str, filename: str, revision: str | None = None) -> Path:
+        return Path(hf_hub_download(repo, filename, revision=revision,
+                                    cache_dir=str(MODELS), token=TOKEN))
+
+    return {
+        "packed": fetch(ENCODER_REPO, ENCODER_FILE),
+        "encoder_dir": Path(snapshot_download(
+            ENCODER_REPO, allow_patterns=["encoder-hf/*"],
+            cache_dir=str(MODELS), token=TOKEN)) / "encoder-hf",
+        "gguf": fetch(DIT_REPO, DIT_FILE),
+        "video_vae": fetch(VAE_REPO, VIDEO_VAE, VAE_REVISION),
+        "audio_vae": fetch(VAE_REPO, AUDIO_VAE, VAE_REVISION),
     }
-    return paths
 
 
 PATHS = bootstrap()

sync.sh

@@ -0,0 +1,31 @@
+#!/usr/bin/env bash
+# Copy the modules the Space needs out of the repo.
+#
+# A Space is a separate git repo, so it cannot import from `../scripts`. The
+# copies are deliberately NOT tracked here: two tracked copies of
+# `ltx_packed_codec.py` would drift, and the one that drifts is always the one
+# nobody is running locally. Run this before uploading.
+set -euo pipefail
+cd "$(dirname "$0")"
+
+MODULES=(ltx_av_generate ltx_av_guide ltx_prompt_encoder ltx_render_clips
+         ltx_conditioning_dump ltx_prompt_embedding_gate ltx_packed_codec
+         ltx_fake_quantize)
+
+mkdir -p scripts src/pacific
+for module in "${MODULES[@]}"; do
+    cp "../scripts/$module.py" scripts/
+done
+cp ../src/pacific/ltx_av.py ../src/pacific/strict_json.py src/pacific/
+
+cat > src/pacific/__init__.py <<'PY'
+"""A deliberately empty `pacific` for the Space.
+
+The real package's `__init__` re-exports the whole conversion toolchain, which
+would drag training, quantization and benchmarking imports into a Space that
+only needs two helpers. Importing `pacific.ltx_av` executes this file, so this
+file has to stay empty for that import to stay cheap.
+"""
+PY
+
+echo "synced ${#MODULES[@]} modules + pacific shim"