Upload GitForCausalLM

config.json

@@ -1,26 +1,34 @@
 {
-  "_name_or_path": "Salesforce/blip-image-captioning-base",
+  "_name_or_path": "microsoft/git-base-coco",
   "architectures": [
-    "BlipForConditionalGeneration"
+    "GitForCausalLM"
   ],
-  "image_text_hidden_size": 256,
-  "initializer_factor": 1.0,
+  "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,
   "initializer_range": 0.02,
-  "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
-  },
+  "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,
   "torch_dtype": "float32",
   "transformers_version": "4.34.0.dev0",
+  "use_cache": true,
   "vision_config": {
     "dropout": 0.0,
     "initializer_factor": 1.0,
-    "initializer_range": 0.02,
-    "model_type": "blip_vision_model",
-    "num_channels": 3
-  }
+    "model_type": "git_vision_model",
+    "projection_dim": 512
+  },
+  "vocab_size": 30522
 }

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 30522,
-  "eos_token_id": 2,
+  "bos_token_id": 101,
+  "eos_token_id": 102,
   "pad_token_id": 0,
   "transformers_version": "4.34.0.dev0"
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1d1bca5f0f0731b4acad13fb03d20e9ff9930a4b56dda55f7f717611f046059d
-size 989816429
+oid sha256:5248453057d697618b5df624a860357fb71095db69b21e7c69ff280e4df6c907
+size 706579985