Hugging Face's logo

zuhri025
/
TTS-urdu-V2

TensorBoard
Safetensors
Model card Files
xet
 Metrics Community
zuhri025 commited on
Commit
1e56c0e
·
verified ·
1 Parent(s): 6c42732

Save checkpoint at step 4500

Browse files
Files changed (8) hide show
  1. checkpoint-4500/checkpoint_info.json +5 -0
  2. checkpoint-4500/config.json +52 -0
  3. checkpoint-4500/generation_config.json +9 -0
  4. checkpoint-4500/model.safetensors +3 -0
  5. checkpoint-4500/tokenizer.json +0 -0
  6. checkpoint-4500/tokenizer_config.json +35 -0
  7. checkpoint-4500/training_config.json +37 -0
  8. checkpoint-4500/training_state.pt +3 -0
checkpoint-4500/checkpoint_info.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "step": 4500,
3
+ "best_val_loss": 6.770168352127075,
4
+ "tensorboard_logs": "tensorboard_logs"
5
+ }
checkpoint-4500/config.json ADDED
@@ -0,0 +1,52 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 3,
8
+ "dtype": "float32",
9
+ "eos_token_id": 3,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 512,
13
+ "initializer_range": 0.041666666666666664,
14
+ "intermediate_size": 2304,
15
+ "layer_types": [
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention"
33
+ ],
34
+ "max_position_embeddings": 1024,
35
+ "max_window_layers": 17,
36
+ "model_type": "qwen3",
37
+ "num_attention_heads": 4,
38
+ "num_hidden_layers": 17,
39
+ "num_key_value_heads": 1,
40
+ "pad_token_id": null,
41
+ "rms_norm_eps": 1e-06,
42
+ "rope_parameters": {
43
+ "rope_theta": 10000,
44
+ "rope_type": "default"
45
+ },
46
+ "sliding_window": null,
47
+ "tie_word_embeddings": false,
48
+ "transformers_version": "5.0.0",
49
+ "use_cache": true,
50
+ "use_sliding_window": false,
51
+ "vocab_size": 23944
52
+ }
checkpoint-4500/generation_config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 3,
4
+ "eos_token_id": 3,
5
+ "output_attentions": false,
6
+ "output_hidden_states": false,
7
+ "transformers_version": "5.0.0",
8
+ "use_cache": true
9
+ }
checkpoint-4500/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60b541bc6a21fa0c74d43f0203902ef6215831b42d6b8f7a86aacd633a6c123c
3
+ size 383397424
checkpoint-4500/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-4500/tokenizer_config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "bos_token": "<|bos|>",
4
+ "eos_token": "<|eos|>",
5
+ "extra_special_tokens": [
6
+ "<|pad|>",
7
+ "<|unk|>",
8
+ "<|bos|>",
9
+ "<|eos|>",
10
+ "<|start_of_speech|>",
11
+ "<|end_of_speech|>",
12
+ "<|start_of_text|>",
13
+ "<|end_of_text|>",
14
+ "<|speech|>",
15
+ "<|sep|>",
16
+ "<|alloy|>",
17
+ "<|amuch|>",
18
+ "<|ash|>",
19
+ "<|ballad|>",
20
+ "<|coral|>",
21
+ "<|dan|>",
22
+ "<|echo|>",
23
+ "<|fable|>",
24
+ "<|nova|>",
25
+ "<|onyx|>",
26
+ "<|sage|>",
27
+ "<|shimmer|>",
28
+ "<|verse|>"
29
+ ],
30
+ "is_local": false,
31
+ "model_max_length": 1000000000000000019884624838656,
32
+ "pad_token": "<|pad|>",
33
+ "tokenizer_class": "TokenizersBackend",
34
+ "unk_token": "<|unk|>"
35
+ }
checkpoint-4500/training_config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "tokenizer_id": "zuhri025/Urdu-Tokenizer",
3
+ "dataset_id": "zuhri025/Urdu-Munch-Processed-s",
4
+ "model_config_id": "ekwek/Soprano-1.1-80M",
5
+ "save_dir": "TTS-urdu-V2",
6
+ "repo_name": "TTS-urdu-V2",
7
+ "device": "cuda",
8
+ "num_workers": 2,
9
+ "pin_memory": true,
10
+ "compile_model": false,
11
+ "use_mixed_precision": true,
12
+ "seed": 1337,
13
+ "max_steps": 20000,
14
+ "batch_size": 8,
15
+ "grad_accum_steps": 2,
16
+ "seq_len": 1024,
17
+ "max_samples": 500000,
18
+ "max_lr": 0.001,
19
+ "min_lr": 5e-05,
20
+ "warmup_ratio": 0.01,
21
+ "betas": [
22
+ 0.9,
23
+ 0.95
24
+ ],
25
+ "weight_decay": 0.1,
26
+ "grad_clip_norm": 1.0,
27
+ "text_loss_weight": 0.5,
28
+ "voice_dropout_prob": 0.1,
29
+ "val_freq": 1500,
30
+ "save_freq": 1500,
31
+ "log_freq": 10,
32
+ "val_steps": 20,
33
+ "test_split_size": 0.01,
34
+ "upload_tensorboard": true,
35
+ "tensorboard_dir": "tensorboard_logs",
36
+ "resume_from_checkpoint": null
37
+ }
checkpoint-4500/training_state.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:092f4e286adfae64127050c5a2e4c92966baba8913867fd0705fe3184ee1039e
3
+ size 766918347