Hugging Face's logo

mlx-community
/
pocket-tts-8bit

Text-to-Speech
Safetensors
MLX
English
mlx-audio
pocket_tts
speech
speech generation
voice cloning
tts
8-bit precision
Model card Files
xet
 Community
prince-canuma commited on
Commit
c90e9b5
·
verified ·
1 Parent(s): 30f894e

Upload folder using huggingface_hub

Browse files
Files changed (4) hide show
  1. README.md +47 -0
  2. config.json +78 -0
  3. model.safetensors +3 -0
  4. model.safetensors.index.json +355 -0
README.md ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ language:
4
+ - en
5
+ base_model:
6
+ - kyutai/pocket-tts
7
+ tags:
8
+ - mlx
9
+ - text-to-speech
10
+ - speech
11
+ - speech generation
12
+ - voice cloning
13
+ - tts
14
+ - mlx-audio
15
+ library_name: mlx-audio
16
+ ---
17
+ # mlx-community/pocket-tts-8bit
18
+
19
+ This model was converted to MLX format from [`mlx-community/pocket-tts`](https://huggingface.co/mlx-community/pocket-tts) using mlx-audio version **0.3.0rc1**.
20
+
21
+ Refer to the [original model card](https://huggingface.co/mlx-community/pocket-tts) for more details on the model.
22
+
23
+ ## Use with mlx-audio
24
+
25
+ ```bash
26
+ pip install -U mlx-audio
27
+ ```
28
+
29
+ ### CLI Example:
30
+ ```bash
31
+ python -m mlx_audio.tts.generate --model mlx-community/pocket-tts-8bit --text "Hello, this is a test."
32
+ ```
33
+
34
+ ### Python Example:
35
+ ```python
36
+ from mlx_audio.tts.utils import load_model
37
+ from mlx_audio.tts.generate import generate_audio
38
+
39
+ model = load_model("mlx-community/pocket-tts-8bit")
40
+ generate_audio(
41
+ model=model,
42
+ text="Hello, this is a test.",
43
+ ref_audio="path_to_audio.wav",
44
+ file_prefix="test_audio",
45
+ )
46
+
47
+ ```
config.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "flow_lm": {
3
+ "dtype": "float32",
4
+ "flow": {
5
+ "dim": 512,
6
+ "depth": 6
7
+ },
8
+ "transformer": {
9
+ "hidden_scale": 4,
10
+ "max_period": 10000,
11
+ "d_model": 1024,
12
+ "num_heads": 16,
13
+ "num_layers": 6
14
+ },
15
+ "lookup_table": {
16
+ "dim": 1024,
17
+ "n_bins": 4000,
18
+ "tokenizer": "sentencepiece",
19
+ "tokenizer_path": "hf://kyutai/pocket-tts-without-voice-cloning/tokenizer.model@d4fdd22ae8c8e1cb3634e150ebeff1dab2d16df3"
20
+ },
21
+ "weights_path": null
22
+ },
23
+ "mimi": {
24
+ "dtype": "float32",
25
+ "sample_rate": 24000,
26
+ "channels": 1,
27
+ "frame_rate": 12.5,
28
+ "seanet": {
29
+ "dimension": 512,
30
+ "channels": 1,
31
+ "n_filters": 64,
32
+ "n_residual_layers": 1,
33
+ "ratios": [
34
+ 6,
35
+ 5,
36
+ 4
37
+ ],
38
+ "kernel_size": 7,
39
+ "residual_kernel_size": 3,
40
+ "last_kernel_size": 3,
41
+ "dilation_base": 2,
42
+ "pad_mode": "constant",
43
+ "compress": 2
44
+ },
45
+ "transformer": {
46
+ "d_model": 512,
47
+ "input_dimension": 512,
48
+ "output_dimensions": [
49
+ 512
50
+ ],
51
+ "num_heads": 8,
52
+ "num_layers": 2,
53
+ "layer_scale": 0.01,
54
+ "context": 250,
55
+ "dim_feedforward": 2048,
56
+ "max_period": 10000.0
57
+ },
58
+ "quantizer": {
59
+ "dimension": 32,
60
+ "output_dimension": 512
61
+ },
62
+ "weights_path": null
63
+ },
64
+ "model_path": null,
65
+ "model_type": "pocket_tts",
66
+ "quantization": {
67
+ "group_size": 64,
68
+ "bits": 4,
69
+ "mode": "affine"
70
+ },
71
+ "quantization_config": {
72
+ "group_size": 64,
73
+ "bits": 4,
74
+ "mode": "affine"
75
+ },
76
+ "weights_path": "hf://kyutai/pocket-tts/tts_b6369a24.safetensors@427e3d61b276ed69fdd03de0d185fa8a8d97fc5b",
77
+ "weights_path_without_voice_cloning": "hf://kyutai/pocket-tts-without-voice-cloning/tts_b6369a24.safetensors@d4fdd22ae8c8e1cb3634e150ebeff1dab2d16df3"
78
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6f9704274f1929a0d09fd3f31ca153a7b6bc4cf4efd8d26408c78dce1813c20
3
+ size 89282784
model.safetensors.index.json ADDED
@@ -0,0 +1,355 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_size": 89241092,
4
+ "total_parameters": 117332322
5
+ },
6
+ "weight_map": {
7
+ "flow_lm.bos_emb": "model.safetensors",
8
+ "flow_lm.conditioner.embed.biases": "model.safetensors",
9
+ "flow_lm.conditioner.embed.scales": "model.safetensors",
10
+ "flow_lm.conditioner.embed.weight": "model.safetensors",
11
+ "flow_lm.emb_mean": "model.safetensors",
12
+ "flow_lm.emb_std": "model.safetensors",
13
+ "flow_lm.flow_net.cond_embed.bias": "model.safetensors",
14
+ "flow_lm.flow_net.cond_embed.biases": "model.safetensors",
15
+ "flow_lm.flow_net.cond_embed.scales": "model.safetensors",
16
+ "flow_lm.flow_net.cond_embed.weight": "model.safetensors",
17
+ "flow_lm.flow_net.final_layer.adaLN_modulation.1.bias": "model.safetensors",
18
+ "flow_lm.flow_net.final_layer.adaLN_modulation.1.biases": "model.safetensors",
19
+ "flow_lm.flow_net.final_layer.adaLN_modulation.1.scales": "model.safetensors",
20
+ "flow_lm.flow_net.final_layer.adaLN_modulation.1.weight": "model.safetensors",
21
+ "flow_lm.flow_net.final_layer.linear.bias": "model.safetensors",
22
+ "flow_lm.flow_net.final_layer.linear.biases": "model.safetensors",
23
+ "flow_lm.flow_net.final_layer.linear.scales": "model.safetensors",
24
+ "flow_lm.flow_net.final_layer.linear.weight": "model.safetensors",
25
+ "flow_lm.flow_net.input_proj.bias": "model.safetensors",
26
+ "flow_lm.flow_net.input_proj.weight": "model.safetensors",
27
+ "flow_lm.flow_net.res_blocks.0.adaLN_modulation.1.bias": "model.safetensors",
28
+ "flow_lm.flow_net.res_blocks.0.adaLN_modulation.1.biases": "model.safetensors",
29
+ "flow_lm.flow_net.res_blocks.0.adaLN_modulation.1.scales": "model.safetensors",
30
+ "flow_lm.flow_net.res_blocks.0.adaLN_modulation.1.weight": "model.safetensors",
31
+ "flow_lm.flow_net.res_blocks.0.in_ln.bias": "model.safetensors",
32
+ "flow_lm.flow_net.res_blocks.0.in_ln.weight": "model.safetensors",
33
+ "flow_lm.flow_net.res_blocks.0.mlp.0.bias": "model.safetensors",
34
+ "flow_lm.flow_net.res_blocks.0.mlp.0.biases": "model.safetensors",
35
+ "flow_lm.flow_net.res_blocks.0.mlp.0.scales": "model.safetensors",
36
+ "flow_lm.flow_net.res_blocks.0.mlp.0.weight": "model.safetensors",
37
+ "flow_lm.flow_net.res_blocks.0.mlp.2.bias": "model.safetensors",
38
+ "flow_lm.flow_net.res_blocks.0.mlp.2.biases": "model.safetensors",
39
+ "flow_lm.flow_net.res_blocks.0.mlp.2.scales": "model.safetensors",
40
+ "flow_lm.flow_net.res_blocks.0.mlp.2.weight": "model.safetensors",
41
+ "flow_lm.flow_net.res_blocks.1.adaLN_modulation.1.bias": "model.safetensors",
42
+ "flow_lm.flow_net.res_blocks.1.adaLN_modulation.1.biases": "model.safetensors",
43
+ "flow_lm.flow_net.res_blocks.1.adaLN_modulation.1.scales": "model.safetensors",
44
+ "flow_lm.flow_net.res_blocks.1.adaLN_modulation.1.weight": "model.safetensors",
45
+ "flow_lm.flow_net.res_blocks.1.in_ln.bias": "model.safetensors",
46
+ "flow_lm.flow_net.res_blocks.1.in_ln.weight": "model.safetensors",
47
+ "flow_lm.flow_net.res_blocks.1.mlp.0.bias": "model.safetensors",
48
+ "flow_lm.flow_net.res_blocks.1.mlp.0.biases": "model.safetensors",
49
+ "flow_lm.flow_net.res_blocks.1.mlp.0.scales": "model.safetensors",
50
+ "flow_lm.flow_net.res_blocks.1.mlp.0.weight": "model.safetensors",
51
+ "flow_lm.flow_net.res_blocks.1.mlp.2.bias": "model.safetensors",
52
+ "flow_lm.flow_net.res_blocks.1.mlp.2.biases": "model.safetensors",
53
+ "flow_lm.flow_net.res_blocks.1.mlp.2.scales": "model.safetensors",
54
+ "flow_lm.flow_net.res_blocks.1.mlp.2.weight": "model.safetensors",
55
+ "flow_lm.flow_net.res_blocks.2.adaLN_modulation.1.bias": "model.safetensors",
56
+ "flow_lm.flow_net.res_blocks.2.adaLN_modulation.1.biases": "model.safetensors",
57
+ "flow_lm.flow_net.res_blocks.2.adaLN_modulation.1.scales": "model.safetensors",
58
+ "flow_lm.flow_net.res_blocks.2.adaLN_modulation.1.weight": "model.safetensors",
59
+ "flow_lm.flow_net.res_blocks.2.in_ln.bias": "model.safetensors",
60
+ "flow_lm.flow_net.res_blocks.2.in_ln.weight": "model.safetensors",
61
+ "flow_lm.flow_net.res_blocks.2.mlp.0.bias": "model.safetensors",
62
+ "flow_lm.flow_net.res_blocks.2.mlp.0.biases": "model.safetensors",
63
+ "flow_lm.flow_net.res_blocks.2.mlp.0.scales": "model.safetensors",
64
+ "flow_lm.flow_net.res_blocks.2.mlp.0.weight": "model.safetensors",
65
+ "flow_lm.flow_net.res_blocks.2.mlp.2.bias": "model.safetensors",
66
+ "flow_lm.flow_net.res_blocks.2.mlp.2.biases": "model.safetensors",
67
+ "flow_lm.flow_net.res_blocks.2.mlp.2.scales": "model.safetensors",
68
+ "flow_lm.flow_net.res_blocks.2.mlp.2.weight": "model.safetensors",
69
+ "flow_lm.flow_net.res_blocks.3.adaLN_modulation.1.bias": "model.safetensors",
70
+ "flow_lm.flow_net.res_blocks.3.adaLN_modulation.1.biases": "model.safetensors",
71
+ "flow_lm.flow_net.res_blocks.3.adaLN_modulation.1.scales": "model.safetensors",
72
+ "flow_lm.flow_net.res_blocks.3.adaLN_modulation.1.weight": "model.safetensors",
73
+ "flow_lm.flow_net.res_blocks.3.in_ln.bias": "model.safetensors",
74
+ "flow_lm.flow_net.res_blocks.3.in_ln.weight": "model.safetensors",
75
+ "flow_lm.flow_net.res_blocks.3.mlp.0.bias": "model.safetensors",
76
+ "flow_lm.flow_net.res_blocks.3.mlp.0.biases": "model.safetensors",
77
+ "flow_lm.flow_net.res_blocks.3.mlp.0.scales": "model.safetensors",
78
+ "flow_lm.flow_net.res_blocks.3.mlp.0.weight": "model.safetensors",
79
+ "flow_lm.flow_net.res_blocks.3.mlp.2.bias": "model.safetensors",
80
+ "flow_lm.flow_net.res_blocks.3.mlp.2.biases": "model.safetensors",
81
+ "flow_lm.flow_net.res_blocks.3.mlp.2.scales": "model.safetensors",
82
+ "flow_lm.flow_net.res_blocks.3.mlp.2.weight": "model.safetensors",
83
+ "flow_lm.flow_net.res_blocks.4.adaLN_modulation.1.bias": "model.safetensors",
84
+ "flow_lm.flow_net.res_blocks.4.adaLN_modulation.1.biases": "model.safetensors",
85
+ "flow_lm.flow_net.res_blocks.4.adaLN_modulation.1.scales": "model.safetensors",
86
+ "flow_lm.flow_net.res_blocks.4.adaLN_modulation.1.weight": "model.safetensors",
87
+ "flow_lm.flow_net.res_blocks.4.in_ln.bias": "model.safetensors",
88
+ "flow_lm.flow_net.res_blocks.4.in_ln.weight": "model.safetensors",
89
+ "flow_lm.flow_net.res_blocks.4.mlp.0.bias": "model.safetensors",
90
+ "flow_lm.flow_net.res_blocks.4.mlp.0.biases": "model.safetensors",
91
+ "flow_lm.flow_net.res_blocks.4.mlp.0.scales": "model.safetensors",
92
+ "flow_lm.flow_net.res_blocks.4.mlp.0.weight": "model.safetensors",
93
+ "flow_lm.flow_net.res_blocks.4.mlp.2.bias": "model.safetensors",
94
+ "flow_lm.flow_net.res_blocks.4.mlp.2.biases": "model.safetensors",
95
+ "flow_lm.flow_net.res_blocks.4.mlp.2.scales": "model.safetensors",
96
+ "flow_lm.flow_net.res_blocks.4.mlp.2.weight": "model.safetensors",
97
+ "flow_lm.flow_net.res_blocks.5.adaLN_modulation.1.bias": "model.safetensors",
98
+ "flow_lm.flow_net.res_blocks.5.adaLN_modulation.1.biases": "model.safetensors",
99
+ "flow_lm.flow_net.res_blocks.5.adaLN_modulation.1.scales": "model.safetensors",
100
+ "flow_lm.flow_net.res_blocks.5.adaLN_modulation.1.weight": "model.safetensors",
101
+ "flow_lm.flow_net.res_blocks.5.in_ln.bias": "model.safetensors",
102
+ "flow_lm.flow_net.res_blocks.5.in_ln.weight": "model.safetensors",
103
+ "flow_lm.flow_net.res_blocks.5.mlp.0.bias": "model.safetensors",
104
+ "flow_lm.flow_net.res_blocks.5.mlp.0.biases": "model.safetensors",
105
+ "flow_lm.flow_net.res_blocks.5.mlp.0.scales": "model.safetensors",
106
+ "flow_lm.flow_net.res_blocks.5.mlp.0.weight": "model.safetensors",
107
+ "flow_lm.flow_net.res_blocks.5.mlp.2.bias": "model.safetensors",
108
+ "flow_lm.flow_net.res_blocks.5.mlp.2.biases": "model.safetensors",
109
+ "flow_lm.flow_net.res_blocks.5.mlp.2.scales": "model.safetensors",
110
+ "flow_lm.flow_net.res_blocks.5.mlp.2.weight": "model.safetensors",
111
+ "flow_lm.flow_net.time_embed.0.freqs": "model.safetensors",
112
+ "flow_lm.flow_net.time_embed.0.mlp.0.bias": "model.safetensors",
113
+ "flow_lm.flow_net.time_embed.0.mlp.0.biases": "model.safetensors",
114
+ "flow_lm.flow_net.time_embed.0.mlp.0.scales": "model.safetensors",
115
+ "flow_lm.flow_net.time_embed.0.mlp.0.weight": "model.safetensors",
116
+ "flow_lm.flow_net.time_embed.0.mlp.2.bias": "model.safetensors",
117
+ "flow_lm.flow_net.time_embed.0.mlp.2.biases": "model.safetensors",
118
+ "flow_lm.flow_net.time_embed.0.mlp.2.scales": "model.safetensors",
119
+ "flow_lm.flow_net.time_embed.0.mlp.2.weight": "model.safetensors",
120
+ "flow_lm.flow_net.time_embed.0.mlp.3.alpha": "model.safetensors",
121
+ "flow_lm.flow_net.time_embed.1.freqs": "model.safetensors",
122
+ "flow_lm.flow_net.time_embed.1.mlp.0.bias": "model.safetensors",
123
+ "flow_lm.flow_net.time_embed.1.mlp.0.biases": "model.safetensors",
124
+ "flow_lm.flow_net.time_embed.1.mlp.0.scales": "model.safetensors",
125
+ "flow_lm.flow_net.time_embed.1.mlp.0.weight": "model.safetensors",
126
+ "flow_lm.flow_net.time_embed.1.mlp.2.bias": "model.safetensors",
127
+ "flow_lm.flow_net.time_embed.1.mlp.2.biases": "model.safetensors",
128
+ "flow_lm.flow_net.time_embed.1.mlp.2.scales": "model.safetensors",
129
+ "flow_lm.flow_net.time_embed.1.mlp.2.weight": "model.safetensors",
130
+ "flow_lm.flow_net.time_embed.1.mlp.3.alpha": "model.safetensors",
131
+ "flow_lm.input_linear.weight": "model.safetensors",
132
+ "flow_lm.out_eos.bias": "model.safetensors",
133
+ "flow_lm.out_eos.biases": "model.safetensors",
134
+ "flow_lm.out_eos.scales": "model.safetensors",
135
+ "flow_lm.out_eos.weight": "model.safetensors",
136
+ "flow_lm.out_norm.bias": "model.safetensors",
137
+ "flow_lm.out_norm.weight": "model.safetensors",
138
+ "flow_lm.transformer.layers.0.linear1.biases": "model.safetensors",
139
+ "flow_lm.transformer.layers.0.linear1.scales": "model.safetensors",
140
+ "flow_lm.transformer.layers.0.linear1.weight": "model.safetensors",
141
+ "flow_lm.transformer.layers.0.linear2.biases": "model.safetensors",
142
+ "flow_lm.transformer.layers.0.linear2.scales": "model.safetensors",
143
+ "flow_lm.transformer.layers.0.linear2.weight": "model.safetensors",
144
+ "flow_lm.transformer.layers.0.norm1.bias": "model.safetensors",
145
+ "flow_lm.transformer.layers.0.norm1.weight": "model.safetensors",
146
+ "flow_lm.transformer.layers.0.norm2.bias": "model.safetensors",
147
+ "flow_lm.transformer.layers.0.norm2.weight": "model.safetensors",
148
+ "flow_lm.transformer.layers.0.self_attn.in_proj.biases": "model.safetensors",
149
+ "flow_lm.transformer.layers.0.self_attn.in_proj.scales": "model.safetensors",
150
+ "flow_lm.transformer.layers.0.self_attn.in_proj.weight": "model.safetensors",
151
+ "flow_lm.transformer.layers.0.self_attn.out_proj.biases": "model.safetensors",
152
+ "flow_lm.transformer.layers.0.self_attn.out_proj.scales": "model.safetensors",
153
+ "flow_lm.transformer.layers.0.self_attn.out_proj.weight": "model.safetensors",
154
+ "flow_lm.transformer.layers.1.linear1.biases": "model.safetensors",
155
+ "flow_lm.transformer.layers.1.linear1.scales": "model.safetensors",
156
+ "flow_lm.transformer.layers.1.linear1.weight": "model.safetensors",
157
+ "flow_lm.transformer.layers.1.linear2.biases": "model.safetensors",
158
+ "flow_lm.transformer.layers.1.linear2.scales": "model.safetensors",
159
+ "flow_lm.transformer.layers.1.linear2.weight": "model.safetensors",
160
+ "flow_lm.transformer.layers.1.norm1.bias": "model.safetensors",
161
+ "flow_lm.transformer.layers.1.norm1.weight": "model.safetensors",
162
+ "flow_lm.transformer.layers.1.norm2.bias": "model.safetensors",
163
+ "flow_lm.transformer.layers.1.norm2.weight": "model.safetensors",
164
+ "flow_lm.transformer.layers.1.self_attn.in_proj.biases": "model.safetensors",
165
+ "flow_lm.transformer.layers.1.self_attn.in_proj.scales": "model.safetensors",
166
+ "flow_lm.transformer.layers.1.self_attn.in_proj.weight": "model.safetensors",
167
+ "flow_lm.transformer.layers.1.self_attn.out_proj.biases": "model.safetensors",
168
+ "flow_lm.transformer.layers.1.self_attn.out_proj.scales": "model.safetensors",
169
+ "flow_lm.transformer.layers.1.self_attn.out_proj.weight": "model.safetensors",
170
+ "flow_lm.transformer.layers.2.linear1.biases": "model.safetensors",
171
+ "flow_lm.transformer.layers.2.linear1.scales": "model.safetensors",
172
+ "flow_lm.transformer.layers.2.linear1.weight": "model.safetensors",
173
+ "flow_lm.transformer.layers.2.linear2.biases": "model.safetensors",
174
+ "flow_lm.transformer.layers.2.linear2.scales": "model.safetensors",
175
+ "flow_lm.transformer.layers.2.linear2.weight": "model.safetensors",
176
+ "flow_lm.transformer.layers.2.norm1.bias": "model.safetensors",
177
+ "flow_lm.transformer.layers.2.norm1.weight": "model.safetensors",
178
+ "flow_lm.transformer.layers.2.norm2.bias": "model.safetensors",
179
+ "flow_lm.transformer.layers.2.norm2.weight": "model.safetensors",
180
+ "flow_lm.transformer.layers.2.self_attn.in_proj.biases": "model.safetensors",
181
+ "flow_lm.transformer.layers.2.self_attn.in_proj.scales": "model.safetensors",
182
+ "flow_lm.transformer.layers.2.self_attn.in_proj.weight": "model.safetensors",
183
+ "flow_lm.transformer.layers.2.self_attn.out_proj.biases": "model.safetensors",
184
+ "flow_lm.transformer.layers.2.self_attn.out_proj.scales": "model.safetensors",
185
+ "flow_lm.transformer.layers.2.self_attn.out_proj.weight": "model.safetensors",
186
+ "flow_lm.transformer.layers.3.linear1.biases": "model.safetensors",
187
+ "flow_lm.transformer.layers.3.linear1.scales": "model.safetensors",
188
+ "flow_lm.transformer.layers.3.linear1.weight": "model.safetensors",
189
+ "flow_lm.transformer.layers.3.linear2.biases": "model.safetensors",
190
+ "flow_lm.transformer.layers.3.linear2.scales": "model.safetensors",
191
+ "flow_lm.transformer.layers.3.linear2.weight": "model.safetensors",
192
+ "flow_lm.transformer.layers.3.norm1.bias": "model.safetensors",
193
+ "flow_lm.transformer.layers.3.norm1.weight": "model.safetensors",
194
+ "flow_lm.transformer.layers.3.norm2.bias": "model.safetensors",
195
+ "flow_lm.transformer.layers.3.norm2.weight": "model.safetensors",
196
+ "flow_lm.transformer.layers.3.self_attn.in_proj.biases": "model.safetensors",
197
+ "flow_lm.transformer.layers.3.self_attn.in_proj.scales": "model.safetensors",
198
+ "flow_lm.transformer.layers.3.self_attn.in_proj.weight": "model.safetensors",
199
+ "flow_lm.transformer.layers.3.self_attn.out_proj.biases": "model.safetensors",
200
+ "flow_lm.transformer.layers.3.self_attn.out_proj.scales": "model.safetensors",
201
+ "flow_lm.transformer.layers.3.self_attn.out_proj.weight": "model.safetensors",
202
+ "flow_lm.transformer.layers.4.linear1.biases": "model.safetensors",
203
+ "flow_lm.transformer.layers.4.linear1.scales": "model.safetensors",
204
+ "flow_lm.transformer.layers.4.linear1.weight": "model.safetensors",
205
+ "flow_lm.transformer.layers.4.linear2.biases": "model.safetensors",
206
+ "flow_lm.transformer.layers.4.linear2.scales": "model.safetensors",
207
+ "flow_lm.transformer.layers.4.linear2.weight": "model.safetensors",
208
+ "flow_lm.transformer.layers.4.norm1.bias": "model.safetensors",
209
+ "flow_lm.transformer.layers.4.norm1.weight": "model.safetensors",
210
+ "flow_lm.transformer.layers.4.norm2.bias": "model.safetensors",
211
+ "flow_lm.transformer.layers.4.norm2.weight": "model.safetensors",
212
+ "flow_lm.transformer.layers.4.self_attn.in_proj.biases": "model.safetensors",
213
+ "flow_lm.transformer.layers.4.self_attn.in_proj.scales": "model.safetensors",
214
+ "flow_lm.transformer.layers.4.self_attn.in_proj.weight": "model.safetensors",
215
+ "flow_lm.transformer.layers.4.self_attn.out_proj.biases": "model.safetensors",
216
+ "flow_lm.transformer.layers.4.self_attn.out_proj.scales": "model.safetensors",
217
+ "flow_lm.transformer.layers.4.self_attn.out_proj.weight": "model.safetensors",
218
+ "flow_lm.transformer.layers.5.linear1.biases": "model.safetensors",
219
+ "flow_lm.transformer.layers.5.linear1.scales": "model.safetensors",
220
+ "flow_lm.transformer.layers.5.linear1.weight": "model.safetensors",
221
+ "flow_lm.transformer.layers.5.linear2.biases": "model.safetensors",
222
+ "flow_lm.transformer.layers.5.linear2.scales": "model.safetensors",
223
+ "flow_lm.transformer.layers.5.linear2.weight": "model.safetensors",
224
+ "flow_lm.transformer.layers.5.norm1.bias": "model.safetensors",
225
+ "flow_lm.transformer.layers.5.norm1.weight": "model.safetensors",
226
+ "flow_lm.transformer.layers.5.norm2.bias": "model.safetensors",
227
+ "flow_lm.transformer.layers.5.norm2.weight": "model.safetensors",
228
+ "flow_lm.transformer.layers.5.self_attn.in_proj.biases": "model.safetensors",
229
+ "flow_lm.transformer.layers.5.self_attn.in_proj.scales": "model.safetensors",
230
+ "flow_lm.transformer.layers.5.self_attn.in_proj.weight": "model.safetensors",
231
+ "flow_lm.transformer.layers.5.self_attn.out_proj.biases": "model.safetensors",
232
+ "flow_lm.transformer.layers.5.self_attn.out_proj.scales": "model.safetensors",
233
+ "flow_lm.transformer.layers.5.self_attn.out_proj.weight": "model.safetensors",
234
+ "mimi.decoder.final_conv1d.conv.conv.bias": "model.safetensors",
235
+ "mimi.decoder.final_conv1d.conv.conv.weight": "model.safetensors",
236
+ "mimi.decoder.init_conv1d.conv.conv.bias": "model.safetensors",
237
+ "mimi.decoder.init_conv1d.conv.conv.weight": "model.safetensors",
238
+ "mimi.decoder.layers.0.residuals.0.block.0.conv.conv.bias": "model.safetensors",
239
+ "mimi.decoder.layers.0.residuals.0.block.0.conv.conv.weight": "model.safetensors",
240
+ "mimi.decoder.layers.0.residuals.0.block.1.conv.conv.bias": "model.safetensors",
241
+ "mimi.decoder.layers.0.residuals.0.block.1.conv.conv.weight": "model.safetensors",
242
+ "mimi.decoder.layers.0.upsample.convtr.convtr.bias": "model.safetensors",
243
+ "mimi.decoder.layers.0.upsample.convtr.convtr.weight": "model.safetensors",
244
+ "mimi.decoder.layers.1.residuals.0.block.0.conv.conv.bias": "model.safetensors",
245
+ "mimi.decoder.layers.1.residuals.0.block.0.conv.conv.weight": "model.safetensors",
246
+ "mimi.decoder.layers.1.residuals.0.block.1.conv.conv.bias": "model.safetensors",
247
+ "mimi.decoder.layers.1.residuals.0.block.1.conv.conv.weight": "model.safetensors",
248
+ "mimi.decoder.layers.1.upsample.convtr.convtr.bias": "model.safetensors",
249
+ "mimi.decoder.layers.1.upsample.convtr.convtr.weight": "model.safetensors",
250
+ "mimi.decoder.layers.2.residuals.0.block.0.conv.conv.bias": "model.safetensors",
251
+ "mimi.decoder.layers.2.residuals.0.block.0.conv.conv.weight": "model.safetensors",
252
+ "mimi.decoder.layers.2.residuals.0.block.1.conv.conv.bias": "model.safetensors",
253
+ "mimi.decoder.layers.2.residuals.0.block.1.conv.conv.weight": "model.safetensors",
254
+ "mimi.decoder.layers.2.upsample.convtr.convtr.bias": "model.safetensors",
255
+ "mimi.decoder.layers.2.upsample.convtr.convtr.weight": "model.safetensors",
256
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear1.biases": "model.safetensors",
257
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear1.scales": "model.safetensors",
258
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear1.weight": "model.safetensors",
259
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear2.biases": "model.safetensors",
260
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear2.scales": "model.safetensors",
261
+ "mimi.decoder_transformer.transformer.layers.0.gating.linear2.weight": "model.safetensors",
262
+ "mimi.decoder_transformer.transformer.layers.0.layer_scale_1.scale": "model.safetensors",
263
+ "mimi.decoder_transformer.transformer.layers.0.layer_scale_2.scale": "model.safetensors",
264
+ "mimi.decoder_transformer.transformer.layers.0.norm1.bias": "model.safetensors",
265
+ "mimi.decoder_transformer.transformer.layers.0.norm1.weight": "model.safetensors",
266
+ "mimi.decoder_transformer.transformer.layers.0.norm2.bias": "model.safetensors",
267
+ "mimi.decoder_transformer.transformer.layers.0.norm2.weight": "model.safetensors",
268
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.in_proj.biases": "model.safetensors",
269
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.in_proj.scales": "model.safetensors",
270
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.in_proj.weight": "model.safetensors",
271
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.out_proj.biases": "model.safetensors",
272
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.out_proj.scales": "model.safetensors",
273
+ "mimi.decoder_transformer.transformer.layers.0.self_attn.out_proj.weight": "model.safetensors",
274
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear1.biases": "model.safetensors",
275
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear1.scales": "model.safetensors",
276
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear1.weight": "model.safetensors",
277
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear2.biases": "model.safetensors",
278
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear2.scales": "model.safetensors",
279
+ "mimi.decoder_transformer.transformer.layers.1.gating.linear2.weight": "model.safetensors",
280
+ "mimi.decoder_transformer.transformer.layers.1.layer_scale_1.scale": "model.safetensors",
281
+ "mimi.decoder_transformer.transformer.layers.1.layer_scale_2.scale": "model.safetensors",
282
+ "mimi.decoder_transformer.transformer.layers.1.norm1.bias": "model.safetensors",
283
+ "mimi.decoder_transformer.transformer.layers.1.norm1.weight": "model.safetensors",
284
+ "mimi.decoder_transformer.transformer.layers.1.norm2.bias": "model.safetensors",
285
+ "mimi.decoder_transformer.transformer.layers.1.norm2.weight": "model.safetensors",
286
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.in_proj.biases": "model.safetensors",
287
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.in_proj.scales": "model.safetensors",
288
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.in_proj.weight": "model.safetensors",
289
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.out_proj.biases": "model.safetensors",
290
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.out_proj.scales": "model.safetensors",
291
+ "mimi.decoder_transformer.transformer.layers.1.self_attn.out_proj.weight": "model.safetensors",
292
+ "mimi.downsample.conv.conv.conv.weight": "model.safetensors",
293
+ "mimi.encoder.final_conv1d.conv.conv.bias": "model.safetensors",
294
+ "mimi.encoder.final_conv1d.conv.conv.weight": "model.safetensors",
295
+ "mimi.encoder.init_conv1d.conv.conv.bias": "model.safetensors",
296
+ "mimi.encoder.init_conv1d.conv.conv.weight": "model.safetensors",
297
+ "mimi.encoder.layers.0.downsample.conv.conv.bias": "model.safetensors",
298
+ "mimi.encoder.layers.0.downsample.conv.conv.weight": "model.safetensors",
299
+ "mimi.encoder.layers.0.residuals.0.block.0.conv.conv.bias": "model.safetensors",
300
+ "mimi.encoder.layers.0.residuals.0.block.0.conv.conv.weight": "model.safetensors",
301
+ "mimi.encoder.layers.0.residuals.0.block.1.conv.conv.bias": "model.safetensors",
302
+ "mimi.encoder.layers.0.residuals.0.block.1.conv.conv.weight": "model.safetensors",
303
+ "mimi.encoder.layers.1.downsample.conv.conv.bias": "model.safetensors",
304
+ "mimi.encoder.layers.1.downsample.conv.conv.weight": "model.safetensors",
305
+ "mimi.encoder.layers.1.residuals.0.block.0.conv.conv.bias": "model.safetensors",
306
+ "mimi.encoder.layers.1.residuals.0.block.0.conv.conv.weight": "model.safetensors",
307
+ "mimi.encoder.layers.1.residuals.0.block.1.conv.conv.bias": "model.safetensors",
308
+ "mimi.encoder.layers.1.residuals.0.block.1.conv.conv.weight": "model.safetensors",
309
+ "mimi.encoder.layers.2.downsample.conv.conv.bias": "model.safetensors",
310
+ "mimi.encoder.layers.2.downsample.conv.conv.weight": "model.safetensors",
311
+ "mimi.encoder.layers.2.residuals.0.block.0.conv.conv.bias": "model.safetensors",
312
+ "mimi.encoder.layers.2.residuals.0.block.0.conv.conv.weight": "model.safetensors",
313
+ "mimi.encoder.layers.2.residuals.0.block.1.conv.conv.bias": "model.safetensors",
314
+ "mimi.encoder.layers.2.residuals.0.block.1.conv.conv.weight": "model.safetensors",
315
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear1.biases": "model.safetensors",
316
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear1.scales": "model.safetensors",
317
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear1.weight": "model.safetensors",
318
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear2.biases": "model.safetensors",
319
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear2.scales": "model.safetensors",
320
+ "mimi.encoder_transformer.transformer.layers.0.gating.linear2.weight": "model.safetensors",
321
+ "mimi.encoder_transformer.transformer.layers.0.layer_scale_1.scale": "model.safetensors",
322
+ "mimi.encoder_transformer.transformer.layers.0.layer_scale_2.scale": "model.safetensors",
323
+ "mimi.encoder_transformer.transformer.layers.0.norm1.bias": "model.safetensors",
324
+ "mimi.encoder_transformer.transformer.layers.0.norm1.weight": "model.safetensors",
325
+ "mimi.encoder_transformer.transformer.layers.0.norm2.bias": "model.safetensors",
326
+ "mimi.encoder_transformer.transformer.layers.0.norm2.weight": "model.safetensors",
327
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.in_proj.biases": "model.safetensors",
328
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.in_proj.scales": "model.safetensors",
329
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.in_proj.weight": "model.safetensors",
330
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.out_proj.biases": "model.safetensors",
331
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.out_proj.scales": "model.safetensors",
332
+ "mimi.encoder_transformer.transformer.layers.0.self_attn.out_proj.weight": "model.safetensors",
333
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear1.biases": "model.safetensors",
334
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear1.scales": "model.safetensors",
335
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear1.weight": "model.safetensors",
336
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear2.biases": "model.safetensors",
337
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear2.scales": "model.safetensors",
338
+ "mimi.encoder_transformer.transformer.layers.1.gating.linear2.weight": "model.safetensors",
339
+ "mimi.encoder_transformer.transformer.layers.1.layer_scale_1.scale": "model.safetensors",
340
+ "mimi.encoder_transformer.transformer.layers.1.layer_scale_2.scale": "model.safetensors",
341
+ "mimi.encoder_transformer.transformer.layers.1.norm1.bias": "model.safetensors",
342
+ "mimi.encoder_transformer.transformer.layers.1.norm1.weight": "model.safetensors",
343
+ "mimi.encoder_transformer.transformer.layers.1.norm2.bias": "model.safetensors",
344
+ "mimi.encoder_transformer.transformer.layers.1.norm2.weight": "model.safetensors",
345
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.in_proj.biases": "model.safetensors",
346
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.in_proj.scales": "model.safetensors",
347
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.in_proj.weight": "model.safetensors",
348
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.out_proj.biases": "model.safetensors",
349
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.out_proj.scales": "model.safetensors",
350
+ "mimi.encoder_transformer.transformer.layers.1.self_attn.out_proj.weight": "model.safetensors",
351
+ "mimi.quantizer.output_proj.weight": "model.safetensors",
352
+ "mimi.upsample.convtr.convtr.convtr.weight": "model.safetensors",
353
+ "speaker_proj_weight": "model.safetensors"
354
+ }
355
+ }