upload ckpt

config.json

@@ -0,0 +1,48 @@
+{
+    "model": {
+      "model_cls": "HawkEye_it",
+      "vit_blip_model_path": "model/VideoChat2/umt_l16_qformer.pth",
+      "llama_model_path": "model/vicuna-7b",
+      "videochat2_model_path": "model/VideoChat2/videochat2_7b_stage2.pth",
+      "freeze_vit": true,
+      "freeze_qformer": false,
+      "max_txt_len": 512,
+      "low_resource": false,
+      "vision_encoder": {
+        "name": "vit_l14",
+        "img_size": 224,
+        "patch_size": 16,
+        "d_model": 1024,
+        "encoder_embed_dim": 1024,
+        "encoder_depth": 24,
+        "encoder_num_heads": 16,
+        "drop_path_rate": 0.0,
+        "num_frames": 32,
+        "tubelet_size": 1,
+        "use_checkpoint": false,
+        "checkpoint_num": 0,
+        "pretrained": "",
+        "return_index": -2,
+        "vit_add_ln": true,
+        "ckpt_num_frame": 4
+      },
+      "num_query_token": 32,
+      "qformer_hidden_dropout_prob": 0.1,
+      "qformer_attention_probs_dropout_prob": 0.1,
+      "qformer_drop_path_rate": 0.2,
+      "extra_num_query_token": 64,
+      "qformer_text_input": true,
+      "system": "",
+      "start_token": "<Video>",
+      "end_token": "</Video>",
+      "img_start_token": "<Image>",
+      "img_end_token": "</Image>",
+      "random_shuffle": true,
+      "use_lora": true,
+      "lora_r": 16,
+      "lora_alpha": 32,
+      "lora_dropout": 0.1
+    },
+    "device": "cuda"
+  }
+  

hawkeye.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5da3b6fd40b214c0572daac77a036e012b5abea3b76a98f0666060628bf382a
+size 775258023