Update config.json
Browse files- config.json +6 -6
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "/
|
| 3 |
"architectures": [
|
| 4 |
"TinyLlavaForConditionalGeneration"
|
| 5 |
],
|
|
@@ -10,7 +10,7 @@
|
|
| 10 |
"ignore_index": -100,
|
| 11 |
"image_aspect_ratio": "square",
|
| 12 |
"image_token_index": -200,
|
| 13 |
-
"llm_model_name_or_path": "/
|
| 14 |
"model_type": "tinyllava",
|
| 15 |
"num_queries": 512,
|
| 16 |
"num_resampler_layers": 3,
|
|
@@ -18,7 +18,7 @@
|
|
| 18 |
"resampler_hidden_size": 768,
|
| 19 |
"text_config": {
|
| 20 |
"_attn_implementation_autoset": true,
|
| 21 |
-
"_name_or_path": "/
|
| 22 |
"architectures": [
|
| 23 |
"Qwen2ForCausalLM"
|
| 24 |
],
|
|
@@ -41,7 +41,7 @@
|
|
| 41 |
"vocab_size": 151936
|
| 42 |
},
|
| 43 |
"tokenizer_model_max_length": 3072,
|
| 44 |
-
"tokenizer_name_or_path": "/
|
| 45 |
"tokenizer_padding_side": "right",
|
| 46 |
"tokenizer_use_fast": false,
|
| 47 |
"torch_dtype": "bfloat16",
|
|
@@ -58,7 +58,7 @@
|
|
| 58 |
"image_size": 384,
|
| 59 |
"intermediate_size": 4304,
|
| 60 |
"layer_norm_eps": 1e-06,
|
| 61 |
-
"model_name_or_path": "/
|
| 62 |
"model_name_or_path2": "",
|
| 63 |
"model_type": "siglip_vision_model",
|
| 64 |
"num_attention_heads": 16,
|
|
@@ -68,7 +68,7 @@
|
|
| 68 |
"vision_feature_layer": -2,
|
| 69 |
"vision_feature_select_strategy": "patch",
|
| 70 |
"vision_hidden_size": 1152,
|
| 71 |
-
"vision_model_name_or_path": "/
|
| 72 |
"vision_model_name_or_path2": "",
|
| 73 |
"vocab_size": 151936
|
| 74 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "Zhang199/TinyLLaVA-Video-Coldstart_NextQA_16",
|
| 3 |
"architectures": [
|
| 4 |
"TinyLlavaForConditionalGeneration"
|
| 5 |
],
|
|
|
|
| 10 |
"ignore_index": -100,
|
| 11 |
"image_aspect_ratio": "square",
|
| 12 |
"image_token_index": -200,
|
| 13 |
+
"llm_model_name_or_path": "Qwen/Qwen2.5-3B",
|
| 14 |
"model_type": "tinyllava",
|
| 15 |
"num_queries": 512,
|
| 16 |
"num_resampler_layers": 3,
|
|
|
|
| 18 |
"resampler_hidden_size": 768,
|
| 19 |
"text_config": {
|
| 20 |
"_attn_implementation_autoset": true,
|
| 21 |
+
"_name_or_path": "Qwen/Qwen2.5-3B",
|
| 22 |
"architectures": [
|
| 23 |
"Qwen2ForCausalLM"
|
| 24 |
],
|
|
|
|
| 41 |
"vocab_size": 151936
|
| 42 |
},
|
| 43 |
"tokenizer_model_max_length": 3072,
|
| 44 |
+
"tokenizer_name_or_path": "Qwen/Qwen2.5-3B",
|
| 45 |
"tokenizer_padding_side": "right",
|
| 46 |
"tokenizer_use_fast": false,
|
| 47 |
"torch_dtype": "bfloat16",
|
|
|
|
| 58 |
"image_size": 384,
|
| 59 |
"intermediate_size": 4304,
|
| 60 |
"layer_norm_eps": 1e-06,
|
| 61 |
+
"model_name_or_path": "google/siglip-so400m-patch14-384",
|
| 62 |
"model_name_or_path2": "",
|
| 63 |
"model_type": "siglip_vision_model",
|
| 64 |
"num_attention_heads": 16,
|
|
|
|
| 68 |
"vision_feature_layer": -2,
|
| 69 |
"vision_feature_select_strategy": "patch",
|
| 70 |
"vision_hidden_size": 1152,
|
| 71 |
+
"vision_model_name_or_path": "google/siglip-so400m-patch14-384",
|
| 72 |
"vision_model_name_or_path2": "",
|
| 73 |
"vocab_size": 151936
|
| 74 |
}
|