Update tiny models for ClapModel
Browse files- config.json +80 -0
- model.safetensors +3 -0
- preprocessor_config.json +21 -0
- tokenizer.json +0 -0
- tokenizer_config.json +22 -0
config.json
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"ClapModel"
|
| 4 |
+
],
|
| 5 |
+
"audio_config": {
|
| 6 |
+
"aff_block_r": 4,
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"attention_probs_dropout_prob": 0.0,
|
| 9 |
+
"depths": [
|
| 10 |
+
2,
|
| 11 |
+
2
|
| 12 |
+
],
|
| 13 |
+
"drop_path_rate": 0.0,
|
| 14 |
+
"dropout": 0.1,
|
| 15 |
+
"enable_fusion": false,
|
| 16 |
+
"enable_patch_layer_norm": true,
|
| 17 |
+
"flatten_patch_embeds": true,
|
| 18 |
+
"freq_ratio": 2,
|
| 19 |
+
"fusion_type": null,
|
| 20 |
+
"hidden_act": "gelu",
|
| 21 |
+
"hidden_dropout_prob": 0.1,
|
| 22 |
+
"hidden_size": 32,
|
| 23 |
+
"image_size": 60,
|
| 24 |
+
"initializer_factor": 1.0,
|
| 25 |
+
"initializer_range": 0.02,
|
| 26 |
+
"intermediate_size": 37,
|
| 27 |
+
"layer_norm_eps": 1e-05,
|
| 28 |
+
"mlp_ratio": 4.0,
|
| 29 |
+
"model_type": "clap_audio_model",
|
| 30 |
+
"num_attention_heads": [
|
| 31 |
+
2,
|
| 32 |
+
2
|
| 33 |
+
],
|
| 34 |
+
"num_channels": 3,
|
| 35 |
+
"num_classes": 527,
|
| 36 |
+
"num_hidden_layers": 2,
|
| 37 |
+
"num_mel_bins": 16,
|
| 38 |
+
"patch_embed_input_channels": 1,
|
| 39 |
+
"patch_embeds_hidden_size": 16,
|
| 40 |
+
"patch_size": 2,
|
| 41 |
+
"patch_stride": 2,
|
| 42 |
+
"projection_dim": 64,
|
| 43 |
+
"projection_hidden_act": "relu",
|
| 44 |
+
"qkv_bias": true,
|
| 45 |
+
"spec_size": 64,
|
| 46 |
+
"window_size": 4
|
| 47 |
+
},
|
| 48 |
+
"dtype": "float32",
|
| 49 |
+
"hidden_size": 32,
|
| 50 |
+
"initializer_factor": 1.0,
|
| 51 |
+
"logit_scale_init_value": 14.285714285714285,
|
| 52 |
+
"model_type": "clap",
|
| 53 |
+
"num_hidden_layers": 4,
|
| 54 |
+
"projection_dim": 64,
|
| 55 |
+
"projection_hidden_act": "relu",
|
| 56 |
+
"text_config": {
|
| 57 |
+
"attention_dropout": 0.1,
|
| 58 |
+
"attention_probs_dropout_prob": 0.1,
|
| 59 |
+
"bos_token_id": 0,
|
| 60 |
+
"dropout": 0.1,
|
| 61 |
+
"eos_token_id": 2,
|
| 62 |
+
"hidden_act": "gelu",
|
| 63 |
+
"hidden_dropout_prob": 0.1,
|
| 64 |
+
"hidden_size": 32,
|
| 65 |
+
"initializer_factor": 1.0,
|
| 66 |
+
"initializer_range": 0.02,
|
| 67 |
+
"intermediate_size": 37,
|
| 68 |
+
"layer_norm_eps": 1e-12,
|
| 69 |
+
"max_position_embeddings": 512,
|
| 70 |
+
"model_type": "clap_text_model",
|
| 71 |
+
"num_attention_heads": 4,
|
| 72 |
+
"num_hidden_layers": 2,
|
| 73 |
+
"pad_token_id": 1,
|
| 74 |
+
"projection_dim": 64,
|
| 75 |
+
"projection_hidden_act": "relu",
|
| 76 |
+
"type_vocab_size": 1,
|
| 77 |
+
"vocab_size": 50265
|
| 78 |
+
},
|
| 79 |
+
"transformers_version": "5.16.0.dev0"
|
| 80 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6438bafd7b1035d5ce2df9693d95006fd04ddb0662a195455a46954d852d15d6
|
| 3 |
+
size 6780776
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length_s": 10,
|
| 3 |
+
"feature_extractor_type": "ClapFeatureExtractor",
|
| 4 |
+
"feature_size": 64,
|
| 5 |
+
"fft_window_size": 1024,
|
| 6 |
+
"frequency_max": 14000,
|
| 7 |
+
"frequency_min": 50,
|
| 8 |
+
"hop_length": 480,
|
| 9 |
+
"max_length_s": 10,
|
| 10 |
+
"n_fft": 1024,
|
| 11 |
+
"nb_frequency_bins": 513,
|
| 12 |
+
"nb_max_frames": 1000,
|
| 13 |
+
"nb_max_samples": 480000,
|
| 14 |
+
"padding": "repeatpad",
|
| 15 |
+
"padding_side": "right",
|
| 16 |
+
"padding_value": 0.0,
|
| 17 |
+
"return_attention_mask": false,
|
| 18 |
+
"sampling_rate": 48000,
|
| 19 |
+
"top_db": null,
|
| 20 |
+
"truncation": "fusion"
|
| 21 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<s>",
|
| 5 |
+
"cls_token": "<s>",
|
| 6 |
+
"eos_token": "</s>",
|
| 7 |
+
"errors": "replace",
|
| 8 |
+
"is_local": true,
|
| 9 |
+
"local_files_only": false,
|
| 10 |
+
"mask_token": "<mask>",
|
| 11 |
+
"max_length": null,
|
| 12 |
+
"model_max_length": 512,
|
| 13 |
+
"pad_to_multiple_of": null,
|
| 14 |
+
"pad_token": "<pad>",
|
| 15 |
+
"pad_token_type_id": 0,
|
| 16 |
+
"padding_side": "right",
|
| 17 |
+
"processor_class": "ClapProcessor",
|
| 18 |
+
"sep_token": "</s>",
|
| 19 |
+
"tokenizer_class": "RobertaTokenizer",
|
| 20 |
+
"trim_offsets": true,
|
| 21 |
+
"unk_token": "<unk>"
|
| 22 |
+
}
|