Training in progress, step 5

config.json

@@ -1,23 +1,23 @@
 {
-  "_name_or_path": "/kaggle/working/violence_finetune-11111",
+  "_name_or_path": "MCG-NJU/videomae-large",
   "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": "Non-Vio",
     "1": "Vio"
   },
   "image_size": 224,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 4096,
   "label2id": {
     "Non-Vio": 0,
     "Vio": 1
@@ -25,10 +25,10 @@
   "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,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3283178833fbc54390b8dc223885dfca27dc697458e48f4a22ee478d12e1311
-size 344937368
+oid sha256:afad76a50dd992041ef25cf9cb147bf090ac8e0608f8fb857af8aef9b54a3c53
+size 1215496248

runs/May21_08-50-43_5e6f8bd3e380/events.out.tfevents.1716281465.5e6f8bd3e380.2466.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fe4faa6c0eb303c87bf47a860e494b946e7e7bcff2aa59daed40b165e83bbc9
+size 7482

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fd4535e8aae2d46275185371b6b4587921d44a8c11dd8171493c8afd9faa4582
+oid sha256:e6f3160f75b9646f19f2b1ec698fddd9d90683151e6fda3de67422ce0fd8ddd8
 size 4920