Upload BlipForConditionalGeneration

config.json

@@ -1,34 +1,26 @@
 {
-  "_name_or_path": "microsoft/git-base",
+  "_name_or_path": "Salesforce/blip-image-captioning-base",
   "architectures": [
-    "GitForCausalLM"
+    "BlipForConditionalGeneration"
   ],
-  "attention_probs_dropout_prob": 0.1,
-  "bos_token_id": 101,
-  "classifier_dropout": null,
-  "eos_token_id": 102,
-  "hidden_act": "gelu",
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "image_text_hidden_size": 256,
+  "initializer_factor": 1.0,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
-  "layer_norm_eps": 1e-12,
-  "max_position_embeddings": 1024,
-  "model_type": "git",
-  "num_attention_heads": 12,
-  "num_hidden_layers": 6,
-  "num_image_with_embedding": null,
-  "pad_token_id": 0,
-  "position_embedding_type": "absolute",
-  "tie_word_embeddings": false,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "blip",
+  "projection_dim": 512,
+  "text_config": {
+    "initializer_factor": 1.0,
+    "model_type": "blip_text_model",
+    "num_attention_heads": 12
+  },
   "torch_dtype": "float32",
   "transformers_version": "4.37.0",
-  "use_cache": true,
   "vision_config": {
     "dropout": 0.0,
     "initializer_factor": 1.0,
-    "model_type": "git_vision_model",
-    "projection_dim": 512
-  },
-  "vocab_size": 30522
+    "initializer_range": 0.02,
+    "model_type": "blip_vision_model",
+    "num_channels": 3
+  }
 }

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 101,
-  "eos_token_id": 102,
+  "bos_token_id": 30522,
+  "eos_token_id": 2,
   "pad_token_id": 0,
   "transformers_version": "4.37.0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:491e38a81ae1c1928b0261d0eac56966b6a81792b8a6a99e82460a02ce8b7aab
-size 706516040
+oid sha256:325e9bb122b05d8c50a4f70676428e0033bc37a6bba4ac202b439b06d6ab0f0d
+size 989717056