patrickvonplaten
commited on
Commit
•
9bdda70
1
Parent(s):
081d555
upload
Browse files- continuous_encoder/config.json +15 -0
- continuous_encoder/diffusion_pytorch_model.bin +3 -0
- decoder/config.json +15 -0
- decoder/diffusion_pytorch_model.bin +3 -0
- melgan/model.onnx +3 -0
- model_index.json +24 -0
- notes_encoder/config.json +15 -0
- notes_encoder/diffusion_pytorch_model.bin +3 -0
- scheduler/scheduler_config.json +13 -0
continuous_encoder/config.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "SpectrogramContEncoder",
|
3 |
+
"_diffusers_version": "0.13.0.dev0",
|
4 |
+
"_name_or_path": "/home/patrick/t5_note/continuous_encoder",
|
5 |
+
"d_ff": 2048,
|
6 |
+
"d_kv": 64,
|
7 |
+
"d_model": 768,
|
8 |
+
"dropout_rate": 0.1,
|
9 |
+
"feed_forward_proj": "gated-gelu",
|
10 |
+
"input_dims": 128,
|
11 |
+
"is_decoder": false,
|
12 |
+
"num_heads": 12,
|
13 |
+
"num_layers": 12,
|
14 |
+
"targets_context_length": 256
|
15 |
+
}
|
continuous_encoder/diffusion_pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a9201a98727653d6e251d666e9037e4ee60852aa3095762bb9e2ada04db24352
|
3 |
+
size 341046195
|
decoder/config.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "T5FilmDecoder",
|
3 |
+
"_diffusers_version": "0.13.0.dev0",
|
4 |
+
"_name_or_path": "/home/patrick/t5_note/decoder",
|
5 |
+
"d_ff": 2048,
|
6 |
+
"d_kv": 64,
|
7 |
+
"d_model": 768,
|
8 |
+
"dropout_rate": 0.1,
|
9 |
+
"feed_forward_proj": "gated-gelu",
|
10 |
+
"input_dims": 128,
|
11 |
+
"max_decoder_noise_time": 20000.0,
|
12 |
+
"num_heads": 12,
|
13 |
+
"num_layers": 12,
|
14 |
+
"targets_length": 256
|
15 |
+
}
|
decoder/diffusion_pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7e1ad13dcace200437032e893d4295cbad6e6d52e2b3e1bd05901038c9f301a4
|
3 |
+
size 954934709
|
melgan/model.onnx
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0f7bfb7c042cfed63b133aa26ec440f7b2d08192823fbc2363499696b3720603
|
3 |
+
size 60487709
|
model_index.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "SpectrogramDiffusionPipeline",
|
3 |
+
"_diffusers_version": "0.13.0.dev0",
|
4 |
+
"continuous_encoder": [
|
5 |
+
"spectrogram_diffusion",
|
6 |
+
"SpectrogramContEncoder"
|
7 |
+
],
|
8 |
+
"decoder": [
|
9 |
+
"diffusers",
|
10 |
+
"T5FilmDecoder"
|
11 |
+
],
|
12 |
+
"melgan": [
|
13 |
+
"diffusers",
|
14 |
+
"OnnxRuntimeModel"
|
15 |
+
],
|
16 |
+
"notes_encoder": [
|
17 |
+
"spectrogram_diffusion",
|
18 |
+
"SpectrogramNotesEncoder"
|
19 |
+
],
|
20 |
+
"scheduler": [
|
21 |
+
"diffusers",
|
22 |
+
"DDPMScheduler"
|
23 |
+
]
|
24 |
+
}
|
notes_encoder/config.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "SpectrogramNotesEncoder",
|
3 |
+
"_diffusers_version": "0.13.0.dev0",
|
4 |
+
"_name_or_path": "/home/patrick/t5_note/notes_encoder",
|
5 |
+
"d_ff": 2048,
|
6 |
+
"d_kv": 64,
|
7 |
+
"d_model": 768,
|
8 |
+
"dropout_rate": 0.1,
|
9 |
+
"feed_forward_proj": "gated-gelu",
|
10 |
+
"is_decoder": false,
|
11 |
+
"max_length": 2048,
|
12 |
+
"num_heads": 12,
|
13 |
+
"num_layers": 12,
|
14 |
+
"vocab_size": 1536
|
15 |
+
}
|
notes_encoder/diffusion_pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5fe0f0e538d111431a1ee7ae0b314e1c58303bfd240fbcdd15185acda131da62
|
3 |
+
size 350876595
|
scheduler/scheduler_config.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_class_name": "DDPMScheduler",
|
3 |
+
"_diffusers_version": "0.13.0.dev0",
|
4 |
+
"beta_end": 0.02,
|
5 |
+
"beta_schedule": "squaredcos_cap_v2",
|
6 |
+
"beta_start": 0.0001,
|
7 |
+
"clip_sample": true,
|
8 |
+
"clip_sample_range": 1.0,
|
9 |
+
"num_train_timesteps": 1000,
|
10 |
+
"prediction_type": "epsilon",
|
11 |
+
"trained_betas": null,
|
12 |
+
"variance_type": "fixed_large"
|
13 |
+
}
|