Hugging Face's logo

yhj137
/
pianist-transformer-rendering

Audio-to-Audio
Safetensors
t5gemma
music
art
piano
midi
Model card Files
xet
 Community
yhj137 commited on
Commit
4450643
·
verified ·
1 Parent(s): 5335437

Upload folder using huggingface_hub

Browse files
Files changed (4) hide show
  1. .DS_Store +0 -0
  2. config.json +125 -0
  3. generation_config.json +7 -0
  4. model.safetensors +3 -0
.DS_Store ADDED
Binary file (6.15 kB). View file
 
config.json ADDED
@@ -0,0 +1,125 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "PianoT5Gemma"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 2,
7
+ "classifier_dropout_rate": 0.0,
8
+ "decoder": {
9
+ "attention_bias": false,
10
+ "attention_dropout": 0.0,
11
+ "attn_logit_softcapping": 50.0,
12
+ "cross_attention_hidden_size": 768,
13
+ "dropout_rate": 0.0,
14
+ "eos_token_id": 3,
15
+ "final_logit_softcapping": 30.0,
16
+ "head_dim": 128,
17
+ "hidden_activation": "gelu_pytorch_tanh",
18
+ "hidden_size": 768,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 3072,
21
+ "is_decoder": true,
22
+ "layer_types": [
23
+ "sliding_attention",
24
+ "full_attention"
25
+ ],
26
+ "max_position_embeddings": 8192,
27
+ "model_type": "t5_gemma_module",
28
+ "num_attention_heads": 8,
29
+ "num_hidden_layers": 2,
30
+ "num_key_value_heads": 4,
31
+ "query_pre_attn_scalar": 256,
32
+ "rms_norm_eps": 1e-06,
33
+ "rope_theta": 10000.0,
34
+ "sliding_window": 4096,
35
+ "use_cache": true,
36
+ "vocab_size": 5389
37
+ },
38
+ "dropout_rate": 0.0,
39
+ "encoder": {
40
+ "attention_bias": false,
41
+ "attention_dropout": 0.0,
42
+ "attn_logit_softcapping": 50.0,
43
+ "dropout_rate": 0.0,
44
+ "eos_token_id": 3,
45
+ "final_logit_softcapping": 30.0,
46
+ "head_dim": 128,
47
+ "hidden_activation": "gelu_pytorch_tanh",
48
+ "hidden_size": 768,
49
+ "initializer_range": 0.02,
50
+ "intermediate_size": 3072,
51
+ "layer_types": [
52
+ "sliding_attention",
53
+ "full_attention",
54
+ "sliding_attention",
55
+ "full_attention",
56
+ "sliding_attention",
57
+ "full_attention",
58
+ "sliding_attention",
59
+ "full_attention",
60
+ "sliding_attention",
61
+ "full_attention"
62
+ ],
63
+ "max_position_embeddings": 8192,
64
+ "model_type": "t5_gemma_module",
65
+ "num_attention_heads": 8,
66
+ "num_hidden_layers": 10,
67
+ "num_key_value_heads": 4,
68
+ "query_pre_attn_scalar": 256,
69
+ "rms_norm_eps": 1e-06,
70
+ "rope_theta": 10000.0,
71
+ "sliding_window": 4096,
72
+ "use_cache": true,
73
+ "vocab_size": 5389
74
+ },
75
+ "eos_token_id": 3,
76
+ "hidden_size": 768,
77
+ "initializer_range": 0.02,
78
+ "is_encoder_decoder": true,
79
+ "mask_token_id": 1,
80
+ "model_type": "t5gemma",
81
+ "pad_token_id": 0,
82
+ "pedal_start": 5261,
83
+ "pitch_start": 5,
84
+ "play_token_id": 4,
85
+ "timing_start": 261,
86
+ "torch_dtype": "bfloat16",
87
+ "transformers_version": "4.54.0",
88
+ "use_cache": true,
89
+ "valid_id_range": [
90
+ [
91
+ 5,
92
+ 133
93
+ ],
94
+ [
95
+ 261,
96
+ 5252
97
+ ],
98
+ [
99
+ 133,
100
+ 261
101
+ ],
102
+ [
103
+ 261,
104
+ 5261
105
+ ],
106
+ [
107
+ 5261,
108
+ 5389
109
+ ],
110
+ [
111
+ 5261,
112
+ 5389
113
+ ],
114
+ [
115
+ 5261,
116
+ 5389
117
+ ],
118
+ [
119
+ 5261,
120
+ 5389
121
+ ]
122
+ ],
123
+ "velocity_start": 133,
124
+ "vocab_size": 5389
125
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 2,
4
+ "eos_token_id": 3,
5
+ "pad_token_id": 0,
6
+ "transformers_version": "4.54.0"
7
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8fb9111efc147adcc61a4885bd42a6e6b625626d9fb8576811916927aa8281d0
3
+ size 271506656