Upload 2 files
Browse files- config.json +140 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,140 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DeSTA25AudioModel"
|
| 4 |
+
],
|
| 5 |
+
"audio_locator": "<|AUDIO|>",
|
| 6 |
+
"connector_mode": "ocar_hybrid",
|
| 7 |
+
"dtype": "bfloat16",
|
| 8 |
+
"encoder_config": {
|
| 9 |
+
"_name_or_path": "openai/whisper-large-v3-turbo",
|
| 10 |
+
"activation_dropout": 0.0,
|
| 11 |
+
"activation_function": "gelu",
|
| 12 |
+
"apply_spec_augment": false,
|
| 13 |
+
"architectures": [
|
| 14 |
+
"WhisperForConditionalGeneration"
|
| 15 |
+
],
|
| 16 |
+
"attention_dropout": 0.0,
|
| 17 |
+
"bos_token_id": 50257,
|
| 18 |
+
"classifier_proj_size": 256,
|
| 19 |
+
"d_model": 1280,
|
| 20 |
+
"decoder_attention_heads": 20,
|
| 21 |
+
"decoder_ffn_dim": 5120,
|
| 22 |
+
"decoder_layerdrop": 0.0,
|
| 23 |
+
"decoder_layers": 4,
|
| 24 |
+
"decoder_start_token_id": 50258,
|
| 25 |
+
"dropout": 0.0,
|
| 26 |
+
"dtype": "float16",
|
| 27 |
+
"encoder_attention_heads": 20,
|
| 28 |
+
"encoder_ffn_dim": 5120,
|
| 29 |
+
"encoder_layerdrop": 0.0,
|
| 30 |
+
"encoder_layers": 32,
|
| 31 |
+
"eos_token_id": 50257,
|
| 32 |
+
"init_std": 0.02,
|
| 33 |
+
"mask_feature_length": 10,
|
| 34 |
+
"mask_feature_min_masks": 0,
|
| 35 |
+
"mask_feature_prob": 0.0,
|
| 36 |
+
"mask_time_length": 10,
|
| 37 |
+
"mask_time_min_masks": 2,
|
| 38 |
+
"mask_time_prob": 0.05,
|
| 39 |
+
"max_source_positions": 1500,
|
| 40 |
+
"max_target_positions": 448,
|
| 41 |
+
"median_filter_width": 7,
|
| 42 |
+
"model_type": "whisper",
|
| 43 |
+
"num_hidden_layers": 32,
|
| 44 |
+
"num_mel_bins": 128,
|
| 45 |
+
"pad_token_id": 50257,
|
| 46 |
+
"scale_embedding": false,
|
| 47 |
+
"use_cache": true,
|
| 48 |
+
"use_weighted_layer_sum": false,
|
| 49 |
+
"vocab_size": 51866
|
| 50 |
+
},
|
| 51 |
+
"encoder_model_id": "openai/whisper-large-v3-turbo",
|
| 52 |
+
"eos_token_id": 151645,
|
| 53 |
+
"info": "\uff24\uff45\uff53\uff54\uff41\uff12\u3002\uff15 \uff21\uff55\uff44\uff49\uff4f",
|
| 54 |
+
"llm_config": {
|
| 55 |
+
"_name_or_path": "Qwen/Qwen3-4B",
|
| 56 |
+
"architectures": [
|
| 57 |
+
"Qwen3ForCausalLM"
|
| 58 |
+
],
|
| 59 |
+
"attention_bias": false,
|
| 60 |
+
"attention_dropout": 0.0,
|
| 61 |
+
"bos_token_id": 151643,
|
| 62 |
+
"dtype": "bfloat16",
|
| 63 |
+
"eos_token_id": 151645,
|
| 64 |
+
"head_dim": 128,
|
| 65 |
+
"hidden_act": "silu",
|
| 66 |
+
"hidden_size": 2560,
|
| 67 |
+
"initializer_range": 0.02,
|
| 68 |
+
"intermediate_size": 9728,
|
| 69 |
+
"layer_types": [
|
| 70 |
+
"full_attention",
|
| 71 |
+
"full_attention",
|
| 72 |
+
"full_attention",
|
| 73 |
+
"full_attention",
|
| 74 |
+
"full_attention",
|
| 75 |
+
"full_attention",
|
| 76 |
+
"full_attention",
|
| 77 |
+
"full_attention",
|
| 78 |
+
"full_attention",
|
| 79 |
+
"full_attention",
|
| 80 |
+
"full_attention",
|
| 81 |
+
"full_attention",
|
| 82 |
+
"full_attention",
|
| 83 |
+
"full_attention",
|
| 84 |
+
"full_attention",
|
| 85 |
+
"full_attention",
|
| 86 |
+
"full_attention",
|
| 87 |
+
"full_attention",
|
| 88 |
+
"full_attention",
|
| 89 |
+
"full_attention",
|
| 90 |
+
"full_attention",
|
| 91 |
+
"full_attention",
|
| 92 |
+
"full_attention",
|
| 93 |
+
"full_attention",
|
| 94 |
+
"full_attention",
|
| 95 |
+
"full_attention",
|
| 96 |
+
"full_attention",
|
| 97 |
+
"full_attention",
|
| 98 |
+
"full_attention",
|
| 99 |
+
"full_attention",
|
| 100 |
+
"full_attention",
|
| 101 |
+
"full_attention",
|
| 102 |
+
"full_attention",
|
| 103 |
+
"full_attention",
|
| 104 |
+
"full_attention",
|
| 105 |
+
"full_attention"
|
| 106 |
+
],
|
| 107 |
+
"max_position_embeddings": 40960,
|
| 108 |
+
"max_window_layers": 36,
|
| 109 |
+
"model_type": "qwen3",
|
| 110 |
+
"num_attention_heads": 32,
|
| 111 |
+
"num_hidden_layers": 36,
|
| 112 |
+
"num_key_value_heads": 8,
|
| 113 |
+
"rms_norm_eps": 1e-06,
|
| 114 |
+
"rope_scaling": null,
|
| 115 |
+
"rope_theta": 1000000,
|
| 116 |
+
"sliding_window": null,
|
| 117 |
+
"tie_word_embeddings": true,
|
| 118 |
+
"use_cache": true,
|
| 119 |
+
"use_sliding_window": false,
|
| 120 |
+
"vocab_size": 151936
|
| 121 |
+
},
|
| 122 |
+
"llm_model_id": "Qwen/Qwen3-4B",
|
| 123 |
+
"model_type": "desta25",
|
| 124 |
+
"ocar_enabled": true,
|
| 125 |
+
"ocar_gate_init": 0.1,
|
| 126 |
+
"ocar_global_num_tokens": 64,
|
| 127 |
+
"ocar_local_downsample": 4,
|
| 128 |
+
"ocar_local_kernel_size": 7,
|
| 129 |
+
"ocar_ortho_diversity_weight": 0.01,
|
| 130 |
+
"ocar_ortho_weight_global": 0.01,
|
| 131 |
+
"ocar_prosody_weight_global": 0.1,
|
| 132 |
+
"ocar_prosody_weight_local": 0.1,
|
| 133 |
+
"pad_token_id": 151645,
|
| 134 |
+
"placeholder_token": "<|video_pad|>",
|
| 135 |
+
"prompt_size": 64,
|
| 136 |
+
"qformer_num_hidden_layers": 6,
|
| 137 |
+
"train_id": 30678,
|
| 138 |
+
"transformers_version": "4.57.3",
|
| 139 |
+
"use_lora": false
|
| 140 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c901a48d5fb5b39eef3333a26af30d81727b842ef9b3b7b1aaf3211a1bced073
|
| 3 |
+
size 2603595872
|