Training in progress, epoch 0

config.json

@@ -1,23 +1,23 @@
 {
-  "_name_or_path": "MCG-NJU/videomae-base",
+  "_name_or_path": "MCG-NJU/videomae-large-finetuned-kinetics",
   "architectures": [
     "VideoMAEForVideoClassification"
   ],
   "attention_probs_dropout_prob": 0.0,
-  "decoder_hidden_size": 384,
-  "decoder_intermediate_size": 1536,
-  "decoder_num_attention_heads": 6,
-  "decoder_num_hidden_layers": 4,
+  "decoder_hidden_size": 512,
+  "decoder_intermediate_size": 2048,
+  "decoder_num_attention_heads": 8,
+  "decoder_num_hidden_layers": 12,
   "hidden_act": "gelu",
   "hidden_dropout_prob": 0.0,
-  "hidden_size": 768,
+  "hidden_size": 1024,
   "id2label": {
     "0": "0",
     "1": "1"
   },
   "image_size": 224,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "0": 0,
     "1": 1
@@ -25,15 +25,15 @@
   "layer_norm_eps": 1e-12,
   "model_type": "videomae",
   "norm_pix_loss": true,
-  "num_attention_heads": 12,
+  "num_attention_heads": 16,
   "num_channels": 3,
   "num_frames": 16,
-  "num_hidden_layers": 12,
+  "num_hidden_layers": 24,
   "patch_size": 16,
   "problem_type": "single_label_classification",
   "qkv_bias": true,
   "torch_dtype": "float32",
   "transformers_version": "4.46.2",
   "tubelet_size": 2,
-  "use_mean_pooling": false
+  "use_mean_pooling": true
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ac71f253208355ebdf2f2078cd4cc008af09da2b14217970f0c3355ce0890afd
-size 344937368
+oid sha256:69f1dbed82493d91e33a78b0ef0aad3449071caa71ac79f564e6cb7e553a4a9d
+size 1215496208

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c9eed42fa7df335fd51cf40373b4d6bceeed3a10af4ee3dcdb543c8487a59235
+oid sha256:c26aa13c0416a4d1cc12c562bd4ab59637ff15b8749cf551cfd1314400b61c93
 size 4859