epoch 1

config.json

@@ -1,7 +1,7 @@
 {
-  "_name_or_path": "Salesforce/blip-vqa-capfilt-large",
+  "_name_or_path": "Salesforce/blip-image-captioning-large",
   "architectures": [
-    "BlipForQuestionAnswering"
+    "BlipForConditionalGeneration"
   ],
   "image_text_hidden_size": 256,
   "initializer_factor": 1.0,
@@ -11,6 +11,7 @@
   "model_type": "blip",
   "projection_dim": 512,
   "text_config": {
+    "encoder_hidden_size": 1024,
     "initializer_factor": 1.0,
     "model_type": "blip_text_model",
     "num_attention_heads": 12
@@ -19,9 +20,13 @@
   "transformers_version": "4.39.3",
   "vision_config": {
     "dropout": 0.0,
+    "hidden_size": 1024,
     "initializer_factor": 1.0,
     "initializer_range": 0.02,
+    "intermediate_size": 4096,
     "model_type": "blip_vision_model",
-    "num_channels": 3
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 24
   }
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e0bcd3494b46756684ddc3c9286ed9c32717ad4d37e9c1d62869add2110e829
-size 1538792184
+oid sha256:86bdefcfb2181f13b351e486bc6508804c8dfd5dce98ec9d0e2efffff325c192
+size 1879010480