update config.json
Browse files- config.json +5 -60
config.json
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"Qwen2_5_VLForConditionalGeneration"
|
| 4 |
],
|
|
@@ -28,79 +29,23 @@
|
|
| 28 |
},
|
| 29 |
"rope_theta": 1000000.0,
|
| 30 |
"sliding_window": 32768,
|
| 31 |
-
"
|
| 32 |
-
"_name_or_path": "UCSC-VLAA/VLAA-Thinker-Qwen2.5VL-3B",
|
| 33 |
-
"architectures": [
|
| 34 |
-
"Qwen2_5_VLForConditionalGeneration"
|
| 35 |
-
],
|
| 36 |
-
"attention_dropout": 0.0,
|
| 37 |
-
"bos_token_id": 151643,
|
| 38 |
-
"eos_token_id": 151645,
|
| 39 |
-
"hidden_act": "silu",
|
| 40 |
-
"hidden_size": 2048,
|
| 41 |
-
"image_token_id": null,
|
| 42 |
-
"initializer_range": 0.02,
|
| 43 |
-
"intermediate_size": 11008,
|
| 44 |
-
"max_position_embeddings": 128000,
|
| 45 |
-
"max_window_layers": 70,
|
| 46 |
-
"model_type": "qwen2_5_vl_text",
|
| 47 |
-
"num_attention_heads": 16,
|
| 48 |
-
"num_hidden_layers": 36,
|
| 49 |
-
"num_key_value_heads": 2,
|
| 50 |
-
"rms_norm_eps": 1e-06,
|
| 51 |
-
"rope_scaling": {
|
| 52 |
-
"mrope_section": [
|
| 53 |
-
16,
|
| 54 |
-
24,
|
| 55 |
-
24
|
| 56 |
-
],
|
| 57 |
-
"rope_type": "default",
|
| 58 |
-
"type": "default"
|
| 59 |
-
},
|
| 60 |
-
"rope_theta": 1000000.0,
|
| 61 |
-
"sliding_window": 32768,
|
| 62 |
-
"tie_word_embeddings": true,
|
| 63 |
-
"torch_dtype": "float16",
|
| 64 |
-
"use_cache": false,
|
| 65 |
-
"use_sliding_window": false,
|
| 66 |
-
"video_token_id": null,
|
| 67 |
-
"vision_end_token_id": 151653,
|
| 68 |
-
"vision_start_token_id": 151652,
|
| 69 |
-
"vision_token_id": 151654,
|
| 70 |
-
"vocab_size": 151936
|
| 71 |
-
},
|
| 72 |
"torch_dtype": "float16",
|
| 73 |
-
"transformers_version": "4.
|
| 74 |
"use_cache": false,
|
| 75 |
"use_sliding_window": false,
|
| 76 |
"video_token_id": 151656,
|
| 77 |
"vision_config": {
|
| 78 |
-
"depth": 32,
|
| 79 |
-
"fullatt_block_indexes": [
|
| 80 |
-
7,
|
| 81 |
-
15,
|
| 82 |
-
23,
|
| 83 |
-
31
|
| 84 |
-
],
|
| 85 |
-
"hidden_act": "silu",
|
| 86 |
"hidden_size": 1280,
|
| 87 |
-
"in_channels": 3,
|
| 88 |
"in_chans": 3,
|
| 89 |
-
"initializer_range": 0.02,
|
| 90 |
-
"intermediate_size": 3420,
|
| 91 |
"model_type": "qwen2_5_vl",
|
| 92 |
-
"num_heads": 16,
|
| 93 |
"out_hidden_size": 2048,
|
| 94 |
-
"patch_size": 14,
|
| 95 |
-
"spatial_merge_size": 2,
|
| 96 |
"spatial_patch_size": 14,
|
| 97 |
-
"temporal_patch_size": 2,
|
| 98 |
"tokens_per_second": 2,
|
| 99 |
-
"torch_dtype": "float16"
|
| 100 |
-
"window_size": 112
|
| 101 |
},
|
| 102 |
"vision_end_token_id": 151653,
|
| 103 |
"vision_start_token_id": 151652,
|
| 104 |
"vision_token_id": 151654,
|
| 105 |
"vocab_size": 151936
|
| 106 |
-
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "UCSC-VLAA/VLAA-Thinker-Qwen2.5VL-3B",
|
| 3 |
"architectures": [
|
| 4 |
"Qwen2_5_VLForConditionalGeneration"
|
| 5 |
],
|
|
|
|
| 29 |
},
|
| 30 |
"rope_theta": 1000000.0,
|
| 31 |
"sliding_window": 32768,
|
| 32 |
+
"tie_word_embeddings": true,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
"torch_dtype": "float16",
|
| 34 |
+
"transformers_version": "4.49.0",
|
| 35 |
"use_cache": false,
|
| 36 |
"use_sliding_window": false,
|
| 37 |
"video_token_id": 151656,
|
| 38 |
"vision_config": {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
"hidden_size": 1280,
|
|
|
|
| 40 |
"in_chans": 3,
|
|
|
|
|
|
|
| 41 |
"model_type": "qwen2_5_vl",
|
|
|
|
| 42 |
"out_hidden_size": 2048,
|
|
|
|
|
|
|
| 43 |
"spatial_patch_size": 14,
|
|
|
|
| 44 |
"tokens_per_second": 2,
|
| 45 |
+
"torch_dtype": "float16"
|
|
|
|
| 46 |
},
|
| 47 |
"vision_end_token_id": 151653,
|
| 48 |
"vision_start_token_id": 151652,
|
| 49 |
"vision_token_id": 151654,
|
| 50 |
"vocab_size": 151936
|
| 51 |
+
}
|