convert to coreml from local ckpt

Author: arda-argmax

README.md

@@ -47,14 +47,17 @@ huggingface-cli login --token YOUR_HF_HUB_TOKEN
 **Step 2:** Prepare the denoise model (MMDiT) Core ML model files (`.mlpackage`)
 
 ```shell
-python -m tests.torch2coreml.test_mmdit --sd3-ckpt-path <path-to-sd3-mmdit.safetensors> --model-version {2b} -o <output-mlpackages-directory> --latent-size {64, 128}
+python -m tests.torch2coreml.test_mmdit --sd3-ckpt-path <path-to-sd3-mmdit.safetensors or model-version-string-from-hub> --model-version {2b} -o <output-mlpackages-directory> --latent-size {64, 128}
 ```
 
 **Step 3:** Prepare the VAE Decoder Core ML model files (`.mlpackage`)
 
 ```shell
-python -m tests.torch2coreml.test_vae --sd3-ckpt-path <path-to-sd3-mmdit.safetensors> -o <output-mlpackages-directory> --latent-size {64, 128}
+python -m tests.torch2coreml.test_vae --sd3-ckpt-path <path-to-sd3-mmdit.safetensors or model-version-string-from-hub> -o <output-mlpackages-directory> --latent-size {64, 128}
 ```
+
+Note:
+- `--sd3-ckpt-path` can be a path to a local `.safetensors` file or a HuggingFace repo (e.g. `stabilityai/stable-diffusion-3-medium`).
 </details>
 
 ## <a name="image-generation-with-python-mlx"></a> Image Generation with Python MLX

tests/torch2coreml/test_mmdit.py

@@ -47,6 +47,7 @@ class TestSD3MMDiT(argmaxtools_test_utils.CoreMLTestsMixin, unittest.TestCase):
 
     @classmethod
     def setUpClass(cls):
+        global TEST_SD3_CKPT_PATH
         cls.model_name = "MultiModalDiffusionTransformer"
         cls.test_output_names = ["denoiser_output"]
         cls.test_cache_dir = TEST_CACHE_DIR
@@ -60,7 +61,7 @@ def setUpClass(cls):
             .eval()
         )
         logger.info("Initialized.")
-        TEST_SD3_CKPT_PATH = hf_hub_download(TEST_SD3_HF_REPO, "sd3_medium.safetensors")
+        TEST_SD3_CKPT_PATH = TEST_SD3_CKPT_PATH or hf_hub_download(TEST_SD3_HF_REPO, "sd3_medium.safetensors")
         if TEST_SD3_CKPT_PATH is not None:
 
             logger.info(f"Loading SD3 model checkpoint from {TEST_SD3_CKPT_PATH}")
@@ -133,6 +134,7 @@ def get_test_inputs(cfg: mmdit.MMDiTConfig) -> Dict[str, torch.Tensor]:
     parser.add_argument("--latent-size", default=TEST_LATENT_SIZE, type=int)
     args = parser.parse_args()
 
+    TEST_SD3_CKPT_PATH = args.sd3_ckpt_path if os.path.exists(args.sd3_ckpt_path) else None
     TEST_SD3_HF_REPO = args.sd3_ckpt_path
     TEST_LATENT_SIZE = args.latent_size
     TEST_CKPT_FILE_NAME = args.ckpt_file_name

tests/torch2coreml/test_vae.py

@@ -45,6 +45,7 @@ class TestSD3VAEDecoder(argmaxtools_test_utils.CoreMLTestsMixin, unittest.TestCa
 
     @classmethod
     def setUpClass(cls):
+        global TEST_SD3_CKPT_PATH
         cls.model_name = "VAEDecoder"
         cls.test_output_names = ["image"]
         cls.test_cache_dir = TEST_CACHE_DIR
@@ -56,7 +57,7 @@ def setUpClass(cls):
         )
         logger.info("Initialized.")
 
-        TEST_SD3_CKPT_PATH = hf_hub_download(TEST_SD3_HF_REPO, "sd3_medium.safetensors")
+        TEST_SD3_CKPT_PATH = TEST_SD3_CKPT_PATH or hf_hub_download(TEST_SD3_HF_REPO, "sd3_medium.safetensors")
         if TEST_SD3_CKPT_PATH is not None:
             logger.info(f"Loading SD3 model checkpoint from {TEST_SD3_CKPT_PATH}")
             _load_vae_decoder_weights(cls.test_torch_model, TEST_SD3_CKPT_PATH)
@@ -103,6 +104,7 @@ def get_test_inputs(config: vae.VAEDecoderConfig) -> Dict[str, torch.Tensor]:
     parser.add_argument("--latent-size", default=TEST_LATENT_SIZE, type=int)
     args = parser.parse_args()
 
+    TEST_SD3_CKPT_PATH = args.sd3_ckpt_path if os.path.exists(args.sd3_ckpt_path) else None
     TEST_SD3_HF_REPO = args.sd3_ckpt_path
     TEST_LATENT_SIZE = args.latent_size