hf-transformers-bot
commited on
Commit
•
50d9e4f
1
Parent(s):
2175441
Upload tiny models for SeamlessM4Tv2Model
Browse files- config.json +124 -0
- generation_config.json +8 -0
- model.safetensors +3 -0
- preprocessor_config.json +10 -0
config.json
ADDED
@@ -0,0 +1,124 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"activation_dropout": 0.0,
|
3 |
+
"activation_function": "relu",
|
4 |
+
"adaptor_dropout": 0.1,
|
5 |
+
"adaptor_kernel_size": 8,
|
6 |
+
"adaptor_stride": 8,
|
7 |
+
"add_adapter": true,
|
8 |
+
"architectures": [
|
9 |
+
"SeamlessM4Tv2Model"
|
10 |
+
],
|
11 |
+
"attention_dropout": 0.1,
|
12 |
+
"attention_probs_dropout_prob": 0.1,
|
13 |
+
"bos_token_id": 2,
|
14 |
+
"char_vocab_size": 4,
|
15 |
+
"conv_depthwise_kernel_size": 31,
|
16 |
+
"decoder_attention_heads": 2,
|
17 |
+
"decoder_ffn_dim": 6,
|
18 |
+
"decoder_layerdrop": 0.05,
|
19 |
+
"decoder_layers": 2,
|
20 |
+
"decoder_start_token_id": 3,
|
21 |
+
"dropout": 0.1,
|
22 |
+
"encoder_attention_heads": 2,
|
23 |
+
"encoder_ffn_dim": 6,
|
24 |
+
"encoder_layerdrop": 0.05,
|
25 |
+
"encoder_layers": 2,
|
26 |
+
"eos_token_id": 3,
|
27 |
+
"feature_projection_input_dim": 160,
|
28 |
+
"hidden_act": "gelu",
|
29 |
+
"hidden_dropout_prob": 0.1,
|
30 |
+
"hidden_size": 6,
|
31 |
+
"initializer_range": 0.02,
|
32 |
+
"is_encoder_decoder": true,
|
33 |
+
"lang_embed_dim": 6,
|
34 |
+
"layer_norm_eps": 1e-05,
|
35 |
+
"leaky_relu_slope": 0.1,
|
36 |
+
"left_max_position_embeddings": 2,
|
37 |
+
"max_new_tokens": null,
|
38 |
+
"max_position_embeddings": 256,
|
39 |
+
"model_type": "seamless_m4t_v2",
|
40 |
+
"num_adapter_layers": 1,
|
41 |
+
"num_attention_heads": 2,
|
42 |
+
"num_conv_pos_embeddings": 8,
|
43 |
+
"num_hidden_layers": 2,
|
44 |
+
"pad_token_id": 0,
|
45 |
+
"position_embeddings_type": "relative_key",
|
46 |
+
"resblock_dilation_sizes": [
|
47 |
+
[
|
48 |
+
1,
|
49 |
+
3,
|
50 |
+
5
|
51 |
+
],
|
52 |
+
[
|
53 |
+
1,
|
54 |
+
3,
|
55 |
+
5
|
56 |
+
],
|
57 |
+
[
|
58 |
+
1,
|
59 |
+
3,
|
60 |
+
5
|
61 |
+
]
|
62 |
+
],
|
63 |
+
"resblock_kernel_sizes": [
|
64 |
+
3,
|
65 |
+
7,
|
66 |
+
11
|
67 |
+
],
|
68 |
+
"right_max_position_embeddings": 1,
|
69 |
+
"sampling_rate": 16000,
|
70 |
+
"scale_embedding": true,
|
71 |
+
"speech_encoder_attention_heads": 2,
|
72 |
+
"speech_encoder_chunk_size": 2,
|
73 |
+
"speech_encoder_dropout": 0.0,
|
74 |
+
"speech_encoder_hidden_act": "swish",
|
75 |
+
"speech_encoder_intermediate_size": 6,
|
76 |
+
"speech_encoder_layerdrop": 0.1,
|
77 |
+
"speech_encoder_layers": 2,
|
78 |
+
"speech_encoder_left_chunk_num": 1,
|
79 |
+
"spkr_embed_dim": 6,
|
80 |
+
"t2u_bos_token_id": 0,
|
81 |
+
"t2u_decoder_attention_heads": 2,
|
82 |
+
"t2u_decoder_ffn_dim": 6,
|
83 |
+
"t2u_decoder_layers": 2,
|
84 |
+
"t2u_encoder_attention_heads": 2,
|
85 |
+
"t2u_encoder_ffn_dim": 6,
|
86 |
+
"t2u_encoder_layers": 2,
|
87 |
+
"t2u_eos_token_id": 2,
|
88 |
+
"t2u_max_position_embeddings": 4096,
|
89 |
+
"t2u_num_langs": 0,
|
90 |
+
"t2u_offset_tgt_lang": 0,
|
91 |
+
"t2u_pad_token_id": 1,
|
92 |
+
"t2u_variance_pred_dropout": 0.5,
|
93 |
+
"t2u_variance_predictor_embed_dim": 6,
|
94 |
+
"t2u_variance_predictor_hidden_dim": 4,
|
95 |
+
"t2u_variance_predictor_kernel_size": 3,
|
96 |
+
"t2u_vocab_size": 20,
|
97 |
+
"torch_dtype": "float32",
|
98 |
+
"transformers_version": "4.36.0.dev0",
|
99 |
+
"unit_embed_dim": 25,
|
100 |
+
"unit_hifi_gan_vocab_size": 20,
|
101 |
+
"unit_hifigan_vocab_vise": 20,
|
102 |
+
"upsample_initial_channel": 32,
|
103 |
+
"upsample_kernel_sizes": [
|
104 |
+
11,
|
105 |
+
8,
|
106 |
+
8,
|
107 |
+
4,
|
108 |
+
4
|
109 |
+
],
|
110 |
+
"upsample_rates": [
|
111 |
+
5,
|
112 |
+
4,
|
113 |
+
4,
|
114 |
+
2,
|
115 |
+
2
|
116 |
+
],
|
117 |
+
"use_cache": true,
|
118 |
+
"var_pred_dropout": 0.5,
|
119 |
+
"variance_predictor_kernel_size": 3,
|
120 |
+
"vocab_size": 20,
|
121 |
+
"vocoder_num_langs": 5,
|
122 |
+
"vocoder_num_spkrs": 5,
|
123 |
+
"vocoder_offset": 0
|
124 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 2,
|
4 |
+
"decoder_start_token_id": 3,
|
5 |
+
"eos_token_id": 3,
|
6 |
+
"pad_token_id": 0,
|
7 |
+
"transformers_version": "4.36.0.dev0"
|
8 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ada293902895a3b8cdc8d8c6a200b623d87da76a8f81f363a533de0063ef196e
|
3 |
+
size 337904
|
preprocessor_config.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"feature_extractor_type": "SeamlessM4TFeatureExtractor",
|
3 |
+
"feature_size": 80,
|
4 |
+
"num_mel_bins": 80,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0.0,
|
7 |
+
"return_attention_mask": true,
|
8 |
+
"sampling_rate": 16000,
|
9 |
+
"stride": 2
|
10 |
+
}
|