fix model repo paths for diffusers from_pretrained()

- Wan: add -Diffusers suffix (Wan-AI/Wan2.1-T2V-1.3B-Diffusers)
- HunyuanVideo: use community repo (hunyuanvideo-community/HunyuanVideo)
- Add e2e generation test script (Wan 1.3B on MPS/CUDA)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: charliewwdev

animatediff/backends/hunyuan.py

@@ -22,7 +22,7 @@
 logger = logging.getLogger(__name__)
 
 HUNYUAN_MODELS = {
-    "default": "tencent/HunyuanVideo",
+    "default": "hunyuanvideo-community/HunyuanVideo",
 }
 
 

animatediff/backends/wan.py

@@ -24,12 +24,12 @@
 logger = logging.getLogger(__name__)
 
 WAN_MODELS = {
-    "1.3B": "Wan-AI/Wan2.1-T2V-1.3B",
-    "14B": "Wan-AI/Wan2.1-T2V-14B",
+    "1.3B": "Wan-AI/Wan2.1-T2V-1.3B-Diffusers",
+    "14B": "Wan-AI/Wan2.1-T2V-14B-Diffusers",
 }
 
 WAN_I2V_MODELS = {
-    "14B": "Wan-AI/Wan2.1-I2V-14B-480P",
+    "14B": "Wan-AI/Wan2.1-I2V-14B-480P-Diffusers",
 }
 
 

tests/test_e2e_generate.py

@@ -0,0 +1,101 @@
+"""
+End-to-end video generation test.
+
+Downloads the smallest available model (Wan 1.3B) and generates a short video
+to verify the full pipeline works: load → generate → save.
+
+This test is slow (downloads ~5GB model on first run) and requires a GPU or
+Apple Silicon with MPS. Mark with pytest -m e2e to run explicitly.
+
+Usage:
+    .venv/bin/python tests/test_e2e_generate.py
+"""
+import os
+import sys
+import time
+
+# Add project root to path
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+
+def test_wan_1_3b_generate():
+    """Generate a short anime clip with Wan 2.1 1.3B."""
+    import torch
+    from animatediff.backends.wan import WanBackend
+
+    # Determine device
+    if torch.cuda.is_available():
+        device = "cuda"
+        dtype = torch.float16
+    elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available():
+        device = "mps"
+        dtype = torch.float32  # MPS works best with float32 for this model
+    else:
+        print("SKIP: No GPU available (need CUDA or MPS)")
+        return
+
+    print(f"\n{'='*60}")
+    print(f"Device: {device} | dtype: {dtype}")
+    print(f"{'='*60}")
+
+    # Load model (will download ~5GB on first run)
+    print("\n[1/3] Loading Wan 2.1 1.3B...")
+    t0 = time.time()
+    backend = WanBackend.load(
+        model_path=None,  # auto: Wan-AI/Wan2.1-T2V-1.3B
+        torch_dtype=dtype,
+        device=device,
+        quantization="none",
+        offload_strategy="model_cpu",  # save memory
+        enable_vae_slicing=True,
+        model_variant="1.3B",
+    )
+    print(f"   Loaded in {time.time() - t0:.1f}s")
+
+    # Generate short video: small resolution, few frames, few steps
+    print("\n[2/3] Generating anime test video...")
+    prompt = "a cute anime girl with blue hair smiling, cherry blossom background, anime style, high quality"
+    t0 = time.time()
+    output = backend.generate(
+        prompt=prompt,
+        negative_prompt="ugly, blurry, low quality",
+        width=480,
+        height=320,
+        num_frames=17,          # ~2 seconds at 8fps
+        num_inference_steps=15,  # draft quality for speed
+        guidance_scale=5.0,
+        seed=42,
+    )
+    gen_time = time.time() - t0
+    print(f"   Generated {len(output.frames)} frames in {gen_time:.1f}s")
+
+    # Save output
+    print("\n[3/3] Saving video...")
+    os.makedirs("samples/e2e_test", exist_ok=True)
+
+    mp4_path = "samples/e2e_test/wan_1.3b_anime.mp4"
+    gif_path = "samples/e2e_test/wan_1.3b_anime.gif"
+    backend.save(output, mp4_path, fps=8)
+    backend.save(output, gif_path, fps=8)
+
+    # Verify
+    assert os.path.exists(mp4_path), f"MP4 not created: {mp4_path}"
+    assert os.path.getsize(mp4_path) > 1000, f"MP4 too small: {os.path.getsize(mp4_path)}"
+    assert os.path.exists(gif_path), f"GIF not created: {gif_path}"
+    assert os.path.getsize(gif_path) > 1000, f"GIF too small: {os.path.getsize(gif_path)}"
+    assert len(output.frames) == 17, f"Expected 17 frames, got {len(output.frames)}"
+    assert output.seed == 42
+    assert output.backend == "wan"
+
+    print(f"\n{'='*60}")
+    print(f"SUCCESS!")
+    print(f"  Frames: {len(output.frames)}")
+    print(f"  MP4: {mp4_path} ({os.path.getsize(mp4_path) / 1024:.0f} KB)")
+    print(f"  GIF: {gif_path} ({os.path.getsize(gif_path) / 1024:.0f} KB)")
+    print(f"  Generation time: {gen_time:.1f}s")
+    print(f"  Prompt: {prompt}")
+    print(f"{'='*60}")
+
+
+if __name__ == "__main__":
+    test_wan_1_3b_generate()