ggcristian
commited on
Commit
•
b9b09c1
1
Parent(s):
e4ce9a7
Upload config.json
Browse files- config.json +76 -0
config.json
ADDED
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"TinyEmoForConditionalGeneration"
|
4 |
+
],
|
5 |
+
"cache_dir": null,
|
6 |
+
"connector_type": "mlp2x_gelu",
|
7 |
+
"hidden_size": 2048,
|
8 |
+
"ignore_index": -100,
|
9 |
+
"image_aspect_ratio": "square",
|
10 |
+
"image_token_index": -200,
|
11 |
+
"llm_model_name_or_path": "apple/TinyLlama-1.1B-Chat-v1.0",
|
12 |
+
"model_type": "tinyemo",
|
13 |
+
"num_queries": 128,
|
14 |
+
"num_resampler_layers": 3,
|
15 |
+
"pad_token": "<|endoftext|>",
|
16 |
+
"pad_token_id": 1,
|
17 |
+
"resampler_hidden_size": 768,
|
18 |
+
"text_config": {
|
19 |
+
"_name_or_path": "apple/TinyLlama-1.1B-Chat-v1.0",
|
20 |
+
"architectures": [
|
21 |
+
"LlamaForCausalLM"
|
22 |
+
],
|
23 |
+
"attention_bias": false,
|
24 |
+
"bos_token_id": 1,
|
25 |
+
"eos_token_id": 2,
|
26 |
+
"hidden_act": "silu",
|
27 |
+
"hidden_size": 2048,
|
28 |
+
"initializer_range": 0.02,
|
29 |
+
"intermediate_size": 5632,
|
30 |
+
"max_position_embeddings": 2048,
|
31 |
+
"model_type": "llama",
|
32 |
+
"num_attention_heads": 32,
|
33 |
+
"num_hidden_layers": 22,
|
34 |
+
"num_key_value_heads": 4,
|
35 |
+
"pretraining_tp": 1,
|
36 |
+
"rms_norm_eps": 1e-05,
|
37 |
+
"rope_scaling": null,
|
38 |
+
"rope_theta": 10000.0,
|
39 |
+
"tie_word_embeddings": false,
|
40 |
+
"torch_dtype": "bfloat16",
|
41 |
+
"transformers_version": "4.35.0",
|
42 |
+
"use_cache": true,
|
43 |
+
"vocab_size": 32000
|
44 |
+
},
|
45 |
+
"tokenizer_model_max_length": 2048,
|
46 |
+
"tokenizer_padding_side": "right",
|
47 |
+
"tokenizer_use_fast": false,
|
48 |
+
"torch_dtype": "bfloat16",
|
49 |
+
"transformers_version": "4.35.0",
|
50 |
+
"tune_type_connector": "full",
|
51 |
+
"tune_type_llm": "full",
|
52 |
+
"tune_type_vision_tower": "frozen",
|
53 |
+
"tune_vision_tower_from_layer": 0,
|
54 |
+
"use_cache": true,
|
55 |
+
"vision_config": {
|
56 |
+
"_name_or_path": "openai/clip-vit-large-patch14",
|
57 |
+
"architectures": [
|
58 |
+
"CLIPModel"
|
59 |
+
],
|
60 |
+
"hidden_act": "quick_gelu",
|
61 |
+
"hidden_size": 1024,
|
62 |
+
"image_size": 224,
|
63 |
+
"intermediate_size": 4096,
|
64 |
+
"layer_norm_eps": 1e-05,
|
65 |
+
"model_type": "clip_vision_model",
|
66 |
+
"num_attention_heads": 16,
|
67 |
+
"num_hidden_layers": 24,
|
68 |
+
"patch_size": 14,
|
69 |
+
"projection_dim": 768
|
70 |
+
},
|
71 |
+
"vision_feature_layer": -2,
|
72 |
+
"vision_feature_select_strategy": "patch",
|
73 |
+
"vision_hidden_size": 1024,
|
74 |
+
"vision_model_name_or_path": "openai/clip-vit-large-patch14",
|
75 |
+
"vocab_size": 32000
|
76 |
+
}
|