Training in progress, epoch 0

config.json

@@ -0,0 +1,30 @@
+{
+  "architectures": [
+    "MoEModelForImageClassification"
+  ],
+  "auto_map": {
+    "AutoConfig": "configuration_moe.MoEConfig",
+    "AutoModelForImageClassification": "modeling_moe.MoEModelForImageClassification"
+  },
+  "baseline_model": "louislu9911/BaseModel-leaf-disease-convnextv2-base-22k-224-0_1_2_3_4",
+  "expert_class_mapping": {
+    "0": [
+      0,
+      4
+    ],
+    "1": [
+      1,
+      2,
+      3
+    ]
+  },
+  "experts": [
+    "louislu9911/Expert1-leaf-disease-convnextv2-base-22k-224-0_4",
+    "louislu9911/Expert2-leaf-disease-convnextv2-base-22k-224-1_2_3"
+  ],
+  "model_type": "moe",
+  "num_classes": 5,
+  "switch_gate": "louislu9911/switch_gate-leaf-disease-convnextv2-base-22k-224",
+  "torch_dtype": "float32",
+  "transformers_version": "4.39.3"
+}

configuration_moe.py

@@ -0,0 +1,38 @@
+from transformers import PretrainedConfig
+from typing import Dict, List
+
+
+MOE_MODEL_TYPE = "moe"
+DEFAULT_HUGGINGFACE_ACCOUNT = "louislu9911"
+model_checkpoint = "convnextv2-base-22k-224"
+
+EXPERTS = [
+    f"{DEFAULT_HUGGINGFACE_ACCOUNT}/Expert1-leaf-disease-{model_checkpoint}-0_4",
+    f"{DEFAULT_HUGGINGFACE_ACCOUNT}/Expert2-leaf-disease-{model_checkpoint}-1_2_3",
+]
+SWITCH_GATE = (
+    f"{DEFAULT_HUGGINGFACE_ACCOUNT}/switch_gate-leaf-disease-{model_checkpoint}"
+)
+BASELINE_MODEL = (
+    f"{DEFAULT_HUGGINGFACE_ACCOUNT}/BaseModel-leaf-disease-{model_checkpoint}-0_1_2_3_4"
+)
+
+
+class MoEConfig(PretrainedConfig):
+    model_type = MOE_MODEL_TYPE
+
+    def __init__(
+        self,
+        experts: List[str] = EXPERTS,
+        switch_gate: str = SWITCH_GATE,
+        baseline_model: str = BASELINE_MODEL,
+        num_classes: int = 5,
+        expert_class_mapping: Dict[int, List[int]] = None,
+        **kwargs,
+    ):
+        self.experts = experts
+        self.switch_gate = switch_gate
+        self.baseline_model = baseline_model
+        self.num_classes = num_classes
+        self.expert_class_mapping = expert_class_mapping
+        super().__init__(**kwargs)

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:345e36ea946911363cbccd406435e8096b6ebc14164b7343361aa85731750ae8
+size 2022035716

preprocessor_config.json

@@ -0,0 +1,37 @@
+{
+  "_valid_processor_keys": [
+    "images",
+    "do_resize",
+    "size",
+    "crop_pct",
+    "resample",
+    "do_rescale",
+    "rescale_factor",
+    "do_normalize",
+    "image_mean",
+    "image_std",
+    "return_tensors",
+    "data_format",
+    "input_data_format"
+  ],
+  "crop_pct": 0.875,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.485,
+    0.456,
+    0.406
+  ],
+  "image_processor_type": "ConvNextImageProcessor",
+  "image_std": [
+    0.229,
+    0.224,
+    0.225
+  ],
+  "resample": 3,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "shortest_edge": 224
+  }
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:caacb8b84ccfb8324214a5a984a2ab7cba2a3dd9141caa0bca618f3fc94aaa8c
+size 4984