mrfakename commited on
Commit
fd8549c
·
verified ·
1 Parent(s): c51bdbd

Upload VibeVoice LAION vocal bursts full finetune artifacts

Browse files
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 3.6373456607845786e+17,
4
+ "train_loss": 1.8379220435955772,
5
+ "train_runtime": 9787.7529,
6
+ "train_samples_per_second": 2.397,
7
+ "train_steps_per_second": 0.15
8
+ }
checkpoint-1400/config.json ADDED
@@ -0,0 +1,148 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": false,
3
+ "acoustic_tokenizer_config": {
4
+ "causal": true,
5
+ "channels": 1,
6
+ "conv_bias": true,
7
+ "conv_norm": "none",
8
+ "corpus_normalize": 0.0,
9
+ "decoder_depths": null,
10
+ "decoder_n_filters": 32,
11
+ "decoder_ratios": [
12
+ 8,
13
+ 5,
14
+ 5,
15
+ 4,
16
+ 2,
17
+ 2
18
+ ],
19
+ "disable_last_norm": true,
20
+ "dtype": "bfloat16",
21
+ "encoder_depths": "3-3-3-3-3-3-8",
22
+ "encoder_n_filters": 32,
23
+ "encoder_ratios": [
24
+ 8,
25
+ 5,
26
+ 5,
27
+ 4,
28
+ 2,
29
+ 2
30
+ ],
31
+ "fix_std": 0.5,
32
+ "layer_scale_init_value": 1e-06,
33
+ "layernorm": "RMSNorm",
34
+ "layernorm_elementwise_affine": true,
35
+ "layernorm_eps": 1e-05,
36
+ "mixer_layer": "depthwise_conv",
37
+ "model_type": "vibevoice_acoustic_tokenizer",
38
+ "pad_mode": "constant",
39
+ "std_dist_type": "gaussian",
40
+ "vae_dim": 64,
41
+ "weight_init_value": 0.01
42
+ },
43
+ "acoustic_vae_dim": 64,
44
+ "architectures": [
45
+ "VibeVoiceASRForConditionalGeneration"
46
+ ],
47
+ "decoder_config": {
48
+ "attention_dropout": 0.0,
49
+ "dtype": "bfloat16",
50
+ "hidden_act": "silu",
51
+ "hidden_size": 3584,
52
+ "initializer_range": 0.02,
53
+ "intermediate_size": 18944,
54
+ "layer_types": [
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention",
72
+ "full_attention",
73
+ "full_attention",
74
+ "full_attention",
75
+ "full_attention",
76
+ "full_attention",
77
+ "full_attention",
78
+ "full_attention",
79
+ "full_attention",
80
+ "full_attention",
81
+ "full_attention",
82
+ "full_attention"
83
+ ],
84
+ "max_position_embeddings": 131072,
85
+ "max_window_layers": 28,
86
+ "model_type": "qwen2",
87
+ "num_attention_heads": 28,
88
+ "num_hidden_layers": 28,
89
+ "num_key_value_heads": 4,
90
+ "rms_norm_eps": 1e-06,
91
+ "rope_scaling": null,
92
+ "rope_theta": 1000000.0,
93
+ "sliding_window": null,
94
+ "use_cache": true,
95
+ "use_mrope": false,
96
+ "use_sliding_window": false,
97
+ "vocab_size": 152064
98
+ },
99
+ "diffusion_head_config": {
100
+ "ddpm_batch_mul": 4,
101
+ "ddpm_beta_schedule": "cosine",
102
+ "ddpm_num_inference_steps": 20,
103
+ "ddpm_num_steps": 1000,
104
+ "diffusion_type": "ddpm",
105
+ "head_ffn_ratio": 3.0,
106
+ "head_layers": 4,
107
+ "hidden_size": 3584,
108
+ "latent_size": 64,
109
+ "model_type": "vibepod_diffusion_head",
110
+ "prediction_type": "v_prediction",
111
+ "rms_norm_eps": 1e-05,
112
+ "speech_vae_dim": 64
113
+ },
114
+ "dtype": "bfloat16",
115
+ "model_type": "vibevoice",
116
+ "semantic_tokenizer_config": {
117
+ "causal": true,
118
+ "channels": 1,
119
+ "conv_bias": true,
120
+ "conv_norm": "none",
121
+ "corpus_normalize": 0.0,
122
+ "disable_last_norm": true,
123
+ "dtype": "bfloat16",
124
+ "encoder_depths": "3-3-3-3-3-3-8",
125
+ "encoder_n_filters": 32,
126
+ "encoder_ratios": [
127
+ 8,
128
+ 5,
129
+ 5,
130
+ 4,
131
+ 2,
132
+ 2
133
+ ],
134
+ "fix_std": 0,
135
+ "layer_scale_init_value": 1e-06,
136
+ "layernorm": "RMSNorm",
137
+ "layernorm_elementwise_affine": true,
138
+ "layernorm_eps": 1e-05,
139
+ "mixer_layer": "depthwise_conv",
140
+ "model_type": "vibevoice_semantic_tokenizer",
141
+ "pad_mode": "constant",
142
+ "std_dist_type": "none",
143
+ "vae_dim": 128,
144
+ "weight_init_value": 0.01
145
+ },
146
+ "semantic_vae_dim": 128,
147
+ "transformers_version": "4.56.2"
148
+ }
checkpoint-1400/generation_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.56.2"
4
+ }
checkpoint-1400/model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5066225664e4cb2ebde1b4f5cfebe15f3b81be3e1deeb14bb8bc15eb963e1eb9
3
+ size 4877662336
checkpoint-1400/model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef5e4885110778e56a053f4f8dfe14aaf996a05ec378f41433caadc9dedec9a4
3
+ size 4932752840
checkpoint-1400/model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2394c7a6f751c4d765a46fd6ed064e4721ddc33769399bcd6ae61e1700192724
3
+ size 4982901128
checkpoint-1400/model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:beb3b5cab2ffa042ceb210be866f2a9573cc4df846043eabc7a11ba3bb2fd48f
3
+ size 2554883402
checkpoint-1400/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd631974c435d355832412b1d7f5ae843dfbf1e77cf56a9724ed54de652225db
3
+ size 30462764371
checkpoint-1400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:deb08623123ecc097619552f4dc66a706e77a96fdff5565dc7333fdd61a749c7
3
+ size 14244
checkpoint-1400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7addc4ba776b2d9d322ad860e0ec3961c6dfd01da1c7abb79bfcc980dbe6b9f3
3
+ size 1064
checkpoint-1400/trainer_state.json ADDED
@@ -0,0 +1,1014 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 0.9546946255807015,
6
+ "eval_steps": 500,
7
+ "global_step": 1400,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.00681924732557644,
14
+ "grad_norm": 38.0,
15
+ "learning_rate": 4.9693251533742335e-05,
16
+ "loss": 7.9038,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.01363849465115288,
21
+ "grad_norm": 13.9375,
22
+ "learning_rate": 4.935241990456715e-05,
23
+ "loss": 3.2871,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.020457741976729317,
28
+ "grad_norm": 14.125,
29
+ "learning_rate": 4.901158827539196e-05,
30
+ "loss": 2.9768,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.02727698930230576,
35
+ "grad_norm": 14.1875,
36
+ "learning_rate": 4.867075664621677e-05,
37
+ "loss": 2.9947,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.0340962366278822,
42
+ "grad_norm": 14.1875,
43
+ "learning_rate": 4.8329925017041586e-05,
44
+ "loss": 2.9812,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.040915483953458634,
49
+ "grad_norm": 13.0,
50
+ "learning_rate": 4.7989093387866394e-05,
51
+ "loss": 2.6322,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.047734731279035075,
56
+ "grad_norm": 16.25,
57
+ "learning_rate": 4.764826175869121e-05,
58
+ "loss": 2.8308,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.05455397860461152,
63
+ "grad_norm": 10.1875,
64
+ "learning_rate": 4.730743012951602e-05,
65
+ "loss": 2.5839,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.06137322593018796,
70
+ "grad_norm": 15.5625,
71
+ "learning_rate": 4.696659850034084e-05,
72
+ "loss": 2.6667,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.0681924732557644,
77
+ "grad_norm": 14.375,
78
+ "learning_rate": 4.6625766871165645e-05,
79
+ "loss": 2.7165,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.07501172058134084,
84
+ "grad_norm": 15.5,
85
+ "learning_rate": 4.628493524199046e-05,
86
+ "loss": 2.7059,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.08183096790691727,
91
+ "grad_norm": 10.5625,
92
+ "learning_rate": 4.594410361281527e-05,
93
+ "loss": 2.483,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.08865021523249371,
98
+ "grad_norm": 14.8125,
99
+ "learning_rate": 4.560327198364008e-05,
100
+ "loss": 2.2635,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.09546946255807015,
105
+ "grad_norm": 13.625,
106
+ "learning_rate": 4.5262440354464896e-05,
107
+ "loss": 2.6928,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.10228870988364659,
112
+ "grad_norm": 11.625,
113
+ "learning_rate": 4.492160872528971e-05,
114
+ "loss": 2.5123,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.10910795720922303,
119
+ "grad_norm": 17.125,
120
+ "learning_rate": 4.4580777096114525e-05,
121
+ "loss": 2.4866,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.11592720453479947,
126
+ "grad_norm": 13.1875,
127
+ "learning_rate": 4.423994546693934e-05,
128
+ "loss": 2.3564,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.12274645186037592,
133
+ "grad_norm": 10.125,
134
+ "learning_rate": 4.389911383776415e-05,
135
+ "loss": 2.1469,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.12956569918595234,
140
+ "grad_norm": 10.8125,
141
+ "learning_rate": 4.3558282208588955e-05,
142
+ "loss": 2.4164,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.1363849465115288,
147
+ "grad_norm": 12.875,
148
+ "learning_rate": 4.321745057941377e-05,
149
+ "loss": 2.5252,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.14320419383710523,
154
+ "grad_norm": 14.375,
155
+ "learning_rate": 4.2876618950238584e-05,
156
+ "loss": 2.3629,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.15002344116268168,
161
+ "grad_norm": 13.5625,
162
+ "learning_rate": 4.25357873210634e-05,
163
+ "loss": 2.2804,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.1568426884882581,
168
+ "grad_norm": 15.25,
169
+ "learning_rate": 4.219495569188821e-05,
170
+ "loss": 2.2377,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.16366193581383454,
175
+ "grad_norm": 10.375,
176
+ "learning_rate": 4.185412406271303e-05,
177
+ "loss": 2.4028,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.170481183139411,
182
+ "grad_norm": 12.3125,
183
+ "learning_rate": 4.1513292433537835e-05,
184
+ "loss": 2.222,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.17730043046498742,
189
+ "grad_norm": 13.4375,
190
+ "learning_rate": 4.117246080436264e-05,
191
+ "loss": 2.4059,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.18411967779056387,
196
+ "grad_norm": 12.5,
197
+ "learning_rate": 4.083162917518746e-05,
198
+ "loss": 2.368,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.1909389251161403,
203
+ "grad_norm": 13.8125,
204
+ "learning_rate": 4.049079754601227e-05,
205
+ "loss": 2.3208,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.19775817244171676,
210
+ "grad_norm": 14.3125,
211
+ "learning_rate": 4.0149965916837086e-05,
212
+ "loss": 2.2395,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.20457741976729318,
217
+ "grad_norm": 11.125,
218
+ "learning_rate": 3.98091342876619e-05,
219
+ "loss": 2.1488,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.2113966670928696,
224
+ "grad_norm": 12.75,
225
+ "learning_rate": 3.946830265848671e-05,
226
+ "loss": 2.0233,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.21821591441844607,
231
+ "grad_norm": 9.3125,
232
+ "learning_rate": 3.912747102931152e-05,
233
+ "loss": 2.2639,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.2250351617440225,
238
+ "grad_norm": 12.875,
239
+ "learning_rate": 3.878663940013633e-05,
240
+ "loss": 2.2124,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.23185440906959895,
245
+ "grad_norm": 9.75,
246
+ "learning_rate": 3.8445807770961145e-05,
247
+ "loss": 2.0552,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.23867365639517538,
252
+ "grad_norm": 16.375,
253
+ "learning_rate": 3.810497614178596e-05,
254
+ "loss": 2.1848,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.24549290372075183,
259
+ "grad_norm": 10.875,
260
+ "learning_rate": 3.7764144512610774e-05,
261
+ "loss": 2.014,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.25231215104632826,
266
+ "grad_norm": 12.0625,
267
+ "learning_rate": 3.742331288343559e-05,
268
+ "loss": 2.1877,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.2591313983719047,
273
+ "grad_norm": 13.5,
274
+ "learning_rate": 3.7082481254260396e-05,
275
+ "loss": 1.9057,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.2659506456974811,
280
+ "grad_norm": 16.25,
281
+ "learning_rate": 3.674164962508521e-05,
282
+ "loss": 2.3365,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.2727698930230576,
287
+ "grad_norm": 8.625,
288
+ "learning_rate": 3.6400817995910025e-05,
289
+ "loss": 2.0162,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.279589140348634,
294
+ "grad_norm": 13.5,
295
+ "learning_rate": 3.605998636673483e-05,
296
+ "loss": 2.1568,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.28640838767421045,
301
+ "grad_norm": 11.3125,
302
+ "learning_rate": 3.571915473755965e-05,
303
+ "loss": 2.0719,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.2932276349997869,
308
+ "grad_norm": 9.75,
309
+ "learning_rate": 3.537832310838446e-05,
310
+ "loss": 1.9952,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.30004688232536336,
315
+ "grad_norm": 9.875,
316
+ "learning_rate": 3.503749147920927e-05,
317
+ "loss": 1.889,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.3068661296509398,
322
+ "grad_norm": 10.5,
323
+ "learning_rate": 3.4696659850034084e-05,
324
+ "loss": 2.0325,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.3136853769765162,
329
+ "grad_norm": 13.0,
330
+ "learning_rate": 3.43558282208589e-05,
331
+ "loss": 2.1277,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.32050462430209264,
336
+ "grad_norm": 10.0,
337
+ "learning_rate": 3.401499659168371e-05,
338
+ "loss": 2.0542,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.32732387162766907,
343
+ "grad_norm": 13.375,
344
+ "learning_rate": 3.367416496250852e-05,
345
+ "loss": 1.9533,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.33414311895324555,
350
+ "grad_norm": 8.8125,
351
+ "learning_rate": 3.3333333333333335e-05,
352
+ "loss": 1.7744,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.340962366278822,
357
+ "grad_norm": 11.375,
358
+ "learning_rate": 3.299250170415815e-05,
359
+ "loss": 1.9968,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.3477816136043984,
364
+ "grad_norm": 10.625,
365
+ "learning_rate": 3.265167007498296e-05,
366
+ "loss": 1.7389,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.35460086092997484,
371
+ "grad_norm": 11.0625,
372
+ "learning_rate": 3.231083844580777e-05,
373
+ "loss": 1.7532,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.3614201082555513,
378
+ "grad_norm": 11.0625,
379
+ "learning_rate": 3.1970006816632586e-05,
380
+ "loss": 1.8736,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.36823935558112775,
385
+ "grad_norm": 12.25,
386
+ "learning_rate": 3.16291751874574e-05,
387
+ "loss": 1.7424,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.3750586029067042,
392
+ "grad_norm": 11.3125,
393
+ "learning_rate": 3.1288343558282215e-05,
394
+ "loss": 1.9172,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.3818778502322806,
399
+ "grad_norm": 10.125,
400
+ "learning_rate": 3.094751192910702e-05,
401
+ "loss": 1.9809,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.38869709755785703,
406
+ "grad_norm": 11.6875,
407
+ "learning_rate": 3.060668029993183e-05,
408
+ "loss": 1.6957,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.3955163448834335,
413
+ "grad_norm": 12.5,
414
+ "learning_rate": 3.026584867075665e-05,
415
+ "loss": 1.9704,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.40233559220900994,
420
+ "grad_norm": 11.0625,
421
+ "learning_rate": 2.992501704158146e-05,
422
+ "loss": 1.9209,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.40915483953458637,
427
+ "grad_norm": 14.875,
428
+ "learning_rate": 2.9584185412406274e-05,
429
+ "loss": 1.8326,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.4159740868601628,
434
+ "grad_norm": 12.6875,
435
+ "learning_rate": 2.9243353783231085e-05,
436
+ "loss": 1.8667,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.4227933341857392,
441
+ "grad_norm": 14.3125,
442
+ "learning_rate": 2.89025221540559e-05,
443
+ "loss": 2.0266,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.4296125815113157,
448
+ "grad_norm": 9.0625,
449
+ "learning_rate": 2.8561690524880707e-05,
450
+ "loss": 1.8784,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.43643182883689213,
455
+ "grad_norm": 9.125,
456
+ "learning_rate": 2.822085889570552e-05,
457
+ "loss": 1.782,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.44325107616246856,
462
+ "grad_norm": 15.4375,
463
+ "learning_rate": 2.7880027266530333e-05,
464
+ "loss": 1.7838,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.450070323488045,
469
+ "grad_norm": 9.625,
470
+ "learning_rate": 2.7539195637355147e-05,
471
+ "loss": 1.8668,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.45688957081362147,
476
+ "grad_norm": 16.75,
477
+ "learning_rate": 2.719836400817996e-05,
478
+ "loss": 1.9149,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.4637088181391979,
483
+ "grad_norm": 11.0,
484
+ "learning_rate": 2.6857532379004773e-05,
485
+ "loss": 1.6853,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.4705280654647743,
490
+ "grad_norm": 10.8125,
491
+ "learning_rate": 2.6516700749829587e-05,
492
+ "loss": 1.7238,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.47734731279035075,
497
+ "grad_norm": 12.625,
498
+ "learning_rate": 2.6175869120654402e-05,
499
+ "loss": 1.7802,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.4841665601159272,
504
+ "grad_norm": 9.1875,
505
+ "learning_rate": 2.583503749147921e-05,
506
+ "loss": 1.7323,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.49098580744150366,
511
+ "grad_norm": 8.9375,
512
+ "learning_rate": 2.549420586230402e-05,
513
+ "loss": 1.7347,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.4978050547670801,
518
+ "grad_norm": 14.5,
519
+ "learning_rate": 2.5153374233128835e-05,
520
+ "loss": 1.9333,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.5046243020926565,
525
+ "grad_norm": 11.8125,
526
+ "learning_rate": 2.481254260395365e-05,
527
+ "loss": 1.799,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.511443549418233,
532
+ "grad_norm": 12.1875,
533
+ "learning_rate": 2.447171097477846e-05,
534
+ "loss": 1.5268,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.5182627967438094,
539
+ "grad_norm": 14.875,
540
+ "learning_rate": 2.4130879345603275e-05,
541
+ "loss": 1.4726,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.5250820440693859,
546
+ "grad_norm": 9.1875,
547
+ "learning_rate": 2.3790047716428086e-05,
548
+ "loss": 1.6679,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.5319012913949622,
553
+ "grad_norm": 11.75,
554
+ "learning_rate": 2.3449216087252897e-05,
555
+ "loss": 1.6178,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.5387205387205387,
560
+ "grad_norm": 11.5625,
561
+ "learning_rate": 2.3108384458077712e-05,
562
+ "loss": 1.6491,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.5455397860461152,
567
+ "grad_norm": 9.125,
568
+ "learning_rate": 2.2767552828902523e-05,
569
+ "loss": 1.6251,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.5523590333716916,
574
+ "grad_norm": 10.875,
575
+ "learning_rate": 2.2426721199727334e-05,
576
+ "loss": 1.7639,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.559178280697268,
581
+ "grad_norm": 12.875,
582
+ "learning_rate": 2.208588957055215e-05,
583
+ "loss": 1.5831,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.5659975280228445,
588
+ "grad_norm": 10.1875,
589
+ "learning_rate": 2.1745057941376963e-05,
590
+ "loss": 1.6586,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.5728167753484209,
595
+ "grad_norm": 11.9375,
596
+ "learning_rate": 2.140422631220177e-05,
597
+ "loss": 1.6956,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.5796360226739974,
602
+ "grad_norm": 8.3125,
603
+ "learning_rate": 2.1063394683026585e-05,
604
+ "loss": 1.6436,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.5864552699995738,
609
+ "grad_norm": 10.125,
610
+ "learning_rate": 2.07225630538514e-05,
611
+ "loss": 1.5014,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.5932745173251502,
616
+ "grad_norm": 14.0625,
617
+ "learning_rate": 2.038173142467621e-05,
618
+ "loss": 1.5563,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.6000937646507267,
623
+ "grad_norm": 12.25,
624
+ "learning_rate": 2.0040899795501022e-05,
625
+ "loss": 1.5834,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.6069130119763031,
630
+ "grad_norm": 11.875,
631
+ "learning_rate": 1.9700068166325836e-05,
632
+ "loss": 1.6049,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.6137322593018796,
637
+ "grad_norm": 10.375,
638
+ "learning_rate": 1.935923653715065e-05,
639
+ "loss": 1.5865,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.620551506627456,
644
+ "grad_norm": 12.0625,
645
+ "learning_rate": 1.9018404907975462e-05,
646
+ "loss": 1.5695,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.6273707539530324,
651
+ "grad_norm": 10.6875,
652
+ "learning_rate": 1.8677573278800273e-05,
653
+ "loss": 1.484,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.6341900012786089,
658
+ "grad_norm": 12.0,
659
+ "learning_rate": 1.8336741649625087e-05,
660
+ "loss": 1.4748,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.6410092486041853,
665
+ "grad_norm": 9.625,
666
+ "learning_rate": 1.79959100204499e-05,
667
+ "loss": 1.4673,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.6478284959297618,
672
+ "grad_norm": 12.3125,
673
+ "learning_rate": 1.7655078391274713e-05,
674
+ "loss": 1.4539,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.6546477432553381,
679
+ "grad_norm": 10.9375,
680
+ "learning_rate": 1.7314246762099524e-05,
681
+ "loss": 1.6066,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.6614669905809146,
686
+ "grad_norm": 12.8125,
687
+ "learning_rate": 1.6973415132924335e-05,
688
+ "loss": 1.6063,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.6682862379064911,
693
+ "grad_norm": 151.0,
694
+ "learning_rate": 1.663258350374915e-05,
695
+ "loss": 1.4019,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.6751054852320675,
700
+ "grad_norm": 10.875,
701
+ "learning_rate": 1.629175187457396e-05,
702
+ "loss": 1.4869,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.681924732557644,
707
+ "grad_norm": 11.375,
708
+ "learning_rate": 1.5950920245398772e-05,
709
+ "loss": 1.5375,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.6887439798832203,
714
+ "grad_norm": 13.3125,
715
+ "learning_rate": 1.5610088616223586e-05,
716
+ "loss": 1.4521,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.6955632272087968,
721
+ "grad_norm": 11.375,
722
+ "learning_rate": 1.52692569870484e-05,
723
+ "loss": 1.5836,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.7023824745343733,
728
+ "grad_norm": 8.6875,
729
+ "learning_rate": 1.492842535787321e-05,
730
+ "loss": 1.4212,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.7092017218599497,
735
+ "grad_norm": 10.125,
736
+ "learning_rate": 1.4587593728698023e-05,
737
+ "loss": 1.5075,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.7160209691855262,
742
+ "grad_norm": 12.4375,
743
+ "learning_rate": 1.4246762099522837e-05,
744
+ "loss": 1.3275,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.7228402165111026,
749
+ "grad_norm": 10.0625,
750
+ "learning_rate": 1.390593047034765e-05,
751
+ "loss": 1.278,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.729659463836679,
756
+ "grad_norm": 14.125,
757
+ "learning_rate": 1.3565098841172461e-05,
758
+ "loss": 1.2336,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.7364787111622555,
763
+ "grad_norm": 9.9375,
764
+ "learning_rate": 1.3224267211997274e-05,
765
+ "loss": 1.3917,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.7432979584878319,
770
+ "grad_norm": 10.25,
771
+ "learning_rate": 1.2883435582822087e-05,
772
+ "loss": 1.3885,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.7501172058134083,
777
+ "grad_norm": 9.9375,
778
+ "learning_rate": 1.25426039536469e-05,
779
+ "loss": 1.3438,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.7569364531389848,
784
+ "grad_norm": 10.1875,
785
+ "learning_rate": 1.2201772324471712e-05,
786
+ "loss": 1.4968,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.7637557004645612,
791
+ "grad_norm": 12.1875,
792
+ "learning_rate": 1.1860940695296524e-05,
793
+ "loss": 1.3124,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.7705749477901377,
798
+ "grad_norm": 9.8125,
799
+ "learning_rate": 1.1520109066121336e-05,
800
+ "loss": 1.282,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.7773941951157141,
805
+ "grad_norm": 12.8125,
806
+ "learning_rate": 1.1179277436946149e-05,
807
+ "loss": 1.2937,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.7842134424412905,
812
+ "grad_norm": 10.5,
813
+ "learning_rate": 1.0838445807770962e-05,
814
+ "loss": 1.3717,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.791032689766867,
819
+ "grad_norm": 10.5625,
820
+ "learning_rate": 1.0497614178595775e-05,
821
+ "loss": 1.2819,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.7978519370924434,
826
+ "grad_norm": 15.75,
827
+ "learning_rate": 1.0156782549420586e-05,
828
+ "loss": 1.3,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.8046711844180199,
833
+ "grad_norm": 9.0,
834
+ "learning_rate": 9.8159509202454e-06,
835
+ "loss": 1.4474,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.8114904317435963,
840
+ "grad_norm": 8.5,
841
+ "learning_rate": 9.475119291070211e-06,
842
+ "loss": 1.412,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.8183096790691727,
847
+ "grad_norm": 10.75,
848
+ "learning_rate": 9.134287661895024e-06,
849
+ "loss": 1.2744,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.8251289263947492,
854
+ "grad_norm": 8.8125,
855
+ "learning_rate": 8.793456032719837e-06,
856
+ "loss": 1.2013,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.8319481737203256,
861
+ "grad_norm": 17.125,
862
+ "learning_rate": 8.45262440354465e-06,
863
+ "loss": 1.463,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.8387674210459021,
868
+ "grad_norm": 11.125,
869
+ "learning_rate": 8.11179277436946e-06,
870
+ "loss": 1.3798,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.8455866683714784,
875
+ "grad_norm": 11.75,
876
+ "learning_rate": 7.770961145194275e-06,
877
+ "loss": 1.2087,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.8524059156970549,
882
+ "grad_norm": 10.125,
883
+ "learning_rate": 7.430129516019086e-06,
884
+ "loss": 1.3037,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.8592251630226314,
889
+ "grad_norm": 9.375,
890
+ "learning_rate": 7.0892978868439e-06,
891
+ "loss": 1.3177,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.8660444103482078,
896
+ "grad_norm": 14.0625,
897
+ "learning_rate": 6.748466257668712e-06,
898
+ "loss": 1.5506,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.8728636576737843,
903
+ "grad_norm": 9.5,
904
+ "learning_rate": 6.4076346284935256e-06,
905
+ "loss": 1.4374,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.8796829049993607,
910
+ "grad_norm": 14.125,
911
+ "learning_rate": 6.066802999318337e-06,
912
+ "loss": 1.2055,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.8865021523249371,
917
+ "grad_norm": 13.75,
918
+ "learning_rate": 5.7259713701431494e-06,
919
+ "loss": 1.4779,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.8933213996505136,
924
+ "grad_norm": 12.1875,
925
+ "learning_rate": 5.385139740967962e-06,
926
+ "loss": 1.3483,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.90014064697609,
931
+ "grad_norm": 12.4375,
932
+ "learning_rate": 5.044308111792775e-06,
933
+ "loss": 1.3771,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.9069598943016665,
938
+ "grad_norm": 9.3125,
939
+ "learning_rate": 4.703476482617587e-06,
940
+ "loss": 1.5074,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.9137791416272429,
945
+ "grad_norm": 9.8125,
946
+ "learning_rate": 4.3626448534424e-06,
947
+ "loss": 1.1781,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.9205983889528193,
952
+ "grad_norm": 9.4375,
953
+ "learning_rate": 4.0218132242672125e-06,
954
+ "loss": 1.3096,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.9274176362783958,
959
+ "grad_norm": 10.0,
960
+ "learning_rate": 3.680981595092025e-06,
961
+ "loss": 1.4104,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.9342368836039722,
966
+ "grad_norm": 10.8125,
967
+ "learning_rate": 3.3401499659168377e-06,
968
+ "loss": 1.3616,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.9410561309295487,
973
+ "grad_norm": 12.4375,
974
+ "learning_rate": 2.9993183367416496e-06,
975
+ "loss": 1.2745,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.9478753782551251,
980
+ "grad_norm": 11.5,
981
+ "learning_rate": 2.6584867075664624e-06,
982
+ "loss": 1.1573,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.9546946255807015,
987
+ "grad_norm": 11.1875,
988
+ "learning_rate": 2.3176550783912747e-06,
989
+ "loss": 1.2449,
990
+ "step": 1400
991
+ }
992
+ ],
993
+ "logging_steps": 10,
994
+ "max_steps": 1467,
995
+ "num_input_tokens_seen": 0,
996
+ "num_train_epochs": 1,
997
+ "save_steps": 100,
998
+ "stateful_callbacks": {
999
+ "TrainerControl": {
1000
+ "args": {
1001
+ "should_epoch_stop": false,
1002
+ "should_evaluate": false,
1003
+ "should_log": false,
1004
+ "should_save": true,
1005
+ "should_training_stop": false
1006
+ },
1007
+ "attributes": {}
1008
+ }
1009
+ },
1010
+ "total_flos": 3.4724378314556685e+17,
1011
+ "train_batch_size": 1,
1012
+ "trial_name": null,
1013
+ "trial_params": null
1014
+ }
checkpoint-1400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbe1c1a78d462e3303123f4ff7facf76d33471dad0aab240d4af4d3300587940
3
+ size 5432
checkpoint-1467/config.json ADDED
@@ -0,0 +1,148 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": false,
3
+ "acoustic_tokenizer_config": {
4
+ "causal": true,
5
+ "channels": 1,
6
+ "conv_bias": true,
7
+ "conv_norm": "none",
8
+ "corpus_normalize": 0.0,
9
+ "decoder_depths": null,
10
+ "decoder_n_filters": 32,
11
+ "decoder_ratios": [
12
+ 8,
13
+ 5,
14
+ 5,
15
+ 4,
16
+ 2,
17
+ 2
18
+ ],
19
+ "disable_last_norm": true,
20
+ "dtype": "bfloat16",
21
+ "encoder_depths": "3-3-3-3-3-3-8",
22
+ "encoder_n_filters": 32,
23
+ "encoder_ratios": [
24
+ 8,
25
+ 5,
26
+ 5,
27
+ 4,
28
+ 2,
29
+ 2
30
+ ],
31
+ "fix_std": 0.5,
32
+ "layer_scale_init_value": 1e-06,
33
+ "layernorm": "RMSNorm",
34
+ "layernorm_elementwise_affine": true,
35
+ "layernorm_eps": 1e-05,
36
+ "mixer_layer": "depthwise_conv",
37
+ "model_type": "vibevoice_acoustic_tokenizer",
38
+ "pad_mode": "constant",
39
+ "std_dist_type": "gaussian",
40
+ "vae_dim": 64,
41
+ "weight_init_value": 0.01
42
+ },
43
+ "acoustic_vae_dim": 64,
44
+ "architectures": [
45
+ "VibeVoiceASRForConditionalGeneration"
46
+ ],
47
+ "decoder_config": {
48
+ "attention_dropout": 0.0,
49
+ "dtype": "bfloat16",
50
+ "hidden_act": "silu",
51
+ "hidden_size": 3584,
52
+ "initializer_range": 0.02,
53
+ "intermediate_size": 18944,
54
+ "layer_types": [
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention",
72
+ "full_attention",
73
+ "full_attention",
74
+ "full_attention",
75
+ "full_attention",
76
+ "full_attention",
77
+ "full_attention",
78
+ "full_attention",
79
+ "full_attention",
80
+ "full_attention",
81
+ "full_attention",
82
+ "full_attention"
83
+ ],
84
+ "max_position_embeddings": 131072,
85
+ "max_window_layers": 28,
86
+ "model_type": "qwen2",
87
+ "num_attention_heads": 28,
88
+ "num_hidden_layers": 28,
89
+ "num_key_value_heads": 4,
90
+ "rms_norm_eps": 1e-06,
91
+ "rope_scaling": null,
92
+ "rope_theta": 1000000.0,
93
+ "sliding_window": null,
94
+ "use_cache": true,
95
+ "use_mrope": false,
96
+ "use_sliding_window": false,
97
+ "vocab_size": 152064
98
+ },
99
+ "diffusion_head_config": {
100
+ "ddpm_batch_mul": 4,
101
+ "ddpm_beta_schedule": "cosine",
102
+ "ddpm_num_inference_steps": 20,
103
+ "ddpm_num_steps": 1000,
104
+ "diffusion_type": "ddpm",
105
+ "head_ffn_ratio": 3.0,
106
+ "head_layers": 4,
107
+ "hidden_size": 3584,
108
+ "latent_size": 64,
109
+ "model_type": "vibepod_diffusion_head",
110
+ "prediction_type": "v_prediction",
111
+ "rms_norm_eps": 1e-05,
112
+ "speech_vae_dim": 64
113
+ },
114
+ "dtype": "bfloat16",
115
+ "model_type": "vibevoice",
116
+ "semantic_tokenizer_config": {
117
+ "causal": true,
118
+ "channels": 1,
119
+ "conv_bias": true,
120
+ "conv_norm": "none",
121
+ "corpus_normalize": 0.0,
122
+ "disable_last_norm": true,
123
+ "dtype": "bfloat16",
124
+ "encoder_depths": "3-3-3-3-3-3-8",
125
+ "encoder_n_filters": 32,
126
+ "encoder_ratios": [
127
+ 8,
128
+ 5,
129
+ 5,
130
+ 4,
131
+ 2,
132
+ 2
133
+ ],
134
+ "fix_std": 0,
135
+ "layer_scale_init_value": 1e-06,
136
+ "layernorm": "RMSNorm",
137
+ "layernorm_elementwise_affine": true,
138
+ "layernorm_eps": 1e-05,
139
+ "mixer_layer": "depthwise_conv",
140
+ "model_type": "vibevoice_semantic_tokenizer",
141
+ "pad_mode": "constant",
142
+ "std_dist_type": "none",
143
+ "vae_dim": 128,
144
+ "weight_init_value": 0.01
145
+ },
146
+ "semantic_vae_dim": 128,
147
+ "transformers_version": "4.56.2"
148
+ }
checkpoint-1467/generation_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.56.2"
4
+ }
checkpoint-1467/model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c70cfe88d16fbd345fe44bdf879c797a7b6ec50e94a024f900cb867adc48e0f
3
+ size 4877662336
checkpoint-1467/model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ee1986d89ea278d2f19e847cf938595ae2590764e199b7dfcd6baced0fd3002
3
+ size 4932752840
checkpoint-1467/model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1733708cf7f6773d071fe61a8364c3ab5cb9a27a9d22fb4ad67511ba061a2d6
3
+ size 4982901128
checkpoint-1467/model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9649879a4c80f30f7bc0509a9662c92c626016f2565d4e91aeb1dbe88287f9bd
3
+ size 2554883402
checkpoint-1467/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1467/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12e9c81ddc5a450a8b3d998f7f3b296d8820ffeb1b0abb367c8ef0733fa9a4e0
3
+ size 30462764371
checkpoint-1467/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:79007e067b58c446e6e09a280a05bcc54f4c5036018dcfb99512be3c5f59229b
3
+ size 14244
checkpoint-1467/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eaf90b9f17744a09a71aca03b154052180851f03a9307650d3ecc38a7ad9b89b
3
+ size 1064
checkpoint-1467/trainer_state.json ADDED
@@ -0,0 +1,1056 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 1.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1467,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.00681924732557644,
14
+ "grad_norm": 38.0,
15
+ "learning_rate": 4.9693251533742335e-05,
16
+ "loss": 7.9038,
17
+ "step": 10
18
+ },
19
+ {
20
+ "epoch": 0.01363849465115288,
21
+ "grad_norm": 13.9375,
22
+ "learning_rate": 4.935241990456715e-05,
23
+ "loss": 3.2871,
24
+ "step": 20
25
+ },
26
+ {
27
+ "epoch": 0.020457741976729317,
28
+ "grad_norm": 14.125,
29
+ "learning_rate": 4.901158827539196e-05,
30
+ "loss": 2.9768,
31
+ "step": 30
32
+ },
33
+ {
34
+ "epoch": 0.02727698930230576,
35
+ "grad_norm": 14.1875,
36
+ "learning_rate": 4.867075664621677e-05,
37
+ "loss": 2.9947,
38
+ "step": 40
39
+ },
40
+ {
41
+ "epoch": 0.0340962366278822,
42
+ "grad_norm": 14.1875,
43
+ "learning_rate": 4.8329925017041586e-05,
44
+ "loss": 2.9812,
45
+ "step": 50
46
+ },
47
+ {
48
+ "epoch": 0.040915483953458634,
49
+ "grad_norm": 13.0,
50
+ "learning_rate": 4.7989093387866394e-05,
51
+ "loss": 2.6322,
52
+ "step": 60
53
+ },
54
+ {
55
+ "epoch": 0.047734731279035075,
56
+ "grad_norm": 16.25,
57
+ "learning_rate": 4.764826175869121e-05,
58
+ "loss": 2.8308,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.05455397860461152,
63
+ "grad_norm": 10.1875,
64
+ "learning_rate": 4.730743012951602e-05,
65
+ "loss": 2.5839,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 0.06137322593018796,
70
+ "grad_norm": 15.5625,
71
+ "learning_rate": 4.696659850034084e-05,
72
+ "loss": 2.6667,
73
+ "step": 90
74
+ },
75
+ {
76
+ "epoch": 0.0681924732557644,
77
+ "grad_norm": 14.375,
78
+ "learning_rate": 4.6625766871165645e-05,
79
+ "loss": 2.7165,
80
+ "step": 100
81
+ },
82
+ {
83
+ "epoch": 0.07501172058134084,
84
+ "grad_norm": 15.5,
85
+ "learning_rate": 4.628493524199046e-05,
86
+ "loss": 2.7059,
87
+ "step": 110
88
+ },
89
+ {
90
+ "epoch": 0.08183096790691727,
91
+ "grad_norm": 10.5625,
92
+ "learning_rate": 4.594410361281527e-05,
93
+ "loss": 2.483,
94
+ "step": 120
95
+ },
96
+ {
97
+ "epoch": 0.08865021523249371,
98
+ "grad_norm": 14.8125,
99
+ "learning_rate": 4.560327198364008e-05,
100
+ "loss": 2.2635,
101
+ "step": 130
102
+ },
103
+ {
104
+ "epoch": 0.09546946255807015,
105
+ "grad_norm": 13.625,
106
+ "learning_rate": 4.5262440354464896e-05,
107
+ "loss": 2.6928,
108
+ "step": 140
109
+ },
110
+ {
111
+ "epoch": 0.10228870988364659,
112
+ "grad_norm": 11.625,
113
+ "learning_rate": 4.492160872528971e-05,
114
+ "loss": 2.5123,
115
+ "step": 150
116
+ },
117
+ {
118
+ "epoch": 0.10910795720922303,
119
+ "grad_norm": 17.125,
120
+ "learning_rate": 4.4580777096114525e-05,
121
+ "loss": 2.4866,
122
+ "step": 160
123
+ },
124
+ {
125
+ "epoch": 0.11592720453479947,
126
+ "grad_norm": 13.1875,
127
+ "learning_rate": 4.423994546693934e-05,
128
+ "loss": 2.3564,
129
+ "step": 170
130
+ },
131
+ {
132
+ "epoch": 0.12274645186037592,
133
+ "grad_norm": 10.125,
134
+ "learning_rate": 4.389911383776415e-05,
135
+ "loss": 2.1469,
136
+ "step": 180
137
+ },
138
+ {
139
+ "epoch": 0.12956569918595234,
140
+ "grad_norm": 10.8125,
141
+ "learning_rate": 4.3558282208588955e-05,
142
+ "loss": 2.4164,
143
+ "step": 190
144
+ },
145
+ {
146
+ "epoch": 0.1363849465115288,
147
+ "grad_norm": 12.875,
148
+ "learning_rate": 4.321745057941377e-05,
149
+ "loss": 2.5252,
150
+ "step": 200
151
+ },
152
+ {
153
+ "epoch": 0.14320419383710523,
154
+ "grad_norm": 14.375,
155
+ "learning_rate": 4.2876618950238584e-05,
156
+ "loss": 2.3629,
157
+ "step": 210
158
+ },
159
+ {
160
+ "epoch": 0.15002344116268168,
161
+ "grad_norm": 13.5625,
162
+ "learning_rate": 4.25357873210634e-05,
163
+ "loss": 2.2804,
164
+ "step": 220
165
+ },
166
+ {
167
+ "epoch": 0.1568426884882581,
168
+ "grad_norm": 15.25,
169
+ "learning_rate": 4.219495569188821e-05,
170
+ "loss": 2.2377,
171
+ "step": 230
172
+ },
173
+ {
174
+ "epoch": 0.16366193581383454,
175
+ "grad_norm": 10.375,
176
+ "learning_rate": 4.185412406271303e-05,
177
+ "loss": 2.4028,
178
+ "step": 240
179
+ },
180
+ {
181
+ "epoch": 0.170481183139411,
182
+ "grad_norm": 12.3125,
183
+ "learning_rate": 4.1513292433537835e-05,
184
+ "loss": 2.222,
185
+ "step": 250
186
+ },
187
+ {
188
+ "epoch": 0.17730043046498742,
189
+ "grad_norm": 13.4375,
190
+ "learning_rate": 4.117246080436264e-05,
191
+ "loss": 2.4059,
192
+ "step": 260
193
+ },
194
+ {
195
+ "epoch": 0.18411967779056387,
196
+ "grad_norm": 12.5,
197
+ "learning_rate": 4.083162917518746e-05,
198
+ "loss": 2.368,
199
+ "step": 270
200
+ },
201
+ {
202
+ "epoch": 0.1909389251161403,
203
+ "grad_norm": 13.8125,
204
+ "learning_rate": 4.049079754601227e-05,
205
+ "loss": 2.3208,
206
+ "step": 280
207
+ },
208
+ {
209
+ "epoch": 0.19775817244171676,
210
+ "grad_norm": 14.3125,
211
+ "learning_rate": 4.0149965916837086e-05,
212
+ "loss": 2.2395,
213
+ "step": 290
214
+ },
215
+ {
216
+ "epoch": 0.20457741976729318,
217
+ "grad_norm": 11.125,
218
+ "learning_rate": 3.98091342876619e-05,
219
+ "loss": 2.1488,
220
+ "step": 300
221
+ },
222
+ {
223
+ "epoch": 0.2113966670928696,
224
+ "grad_norm": 12.75,
225
+ "learning_rate": 3.946830265848671e-05,
226
+ "loss": 2.0233,
227
+ "step": 310
228
+ },
229
+ {
230
+ "epoch": 0.21821591441844607,
231
+ "grad_norm": 9.3125,
232
+ "learning_rate": 3.912747102931152e-05,
233
+ "loss": 2.2639,
234
+ "step": 320
235
+ },
236
+ {
237
+ "epoch": 0.2250351617440225,
238
+ "grad_norm": 12.875,
239
+ "learning_rate": 3.878663940013633e-05,
240
+ "loss": 2.2124,
241
+ "step": 330
242
+ },
243
+ {
244
+ "epoch": 0.23185440906959895,
245
+ "grad_norm": 9.75,
246
+ "learning_rate": 3.8445807770961145e-05,
247
+ "loss": 2.0552,
248
+ "step": 340
249
+ },
250
+ {
251
+ "epoch": 0.23867365639517538,
252
+ "grad_norm": 16.375,
253
+ "learning_rate": 3.810497614178596e-05,
254
+ "loss": 2.1848,
255
+ "step": 350
256
+ },
257
+ {
258
+ "epoch": 0.24549290372075183,
259
+ "grad_norm": 10.875,
260
+ "learning_rate": 3.7764144512610774e-05,
261
+ "loss": 2.014,
262
+ "step": 360
263
+ },
264
+ {
265
+ "epoch": 0.25231215104632826,
266
+ "grad_norm": 12.0625,
267
+ "learning_rate": 3.742331288343559e-05,
268
+ "loss": 2.1877,
269
+ "step": 370
270
+ },
271
+ {
272
+ "epoch": 0.2591313983719047,
273
+ "grad_norm": 13.5,
274
+ "learning_rate": 3.7082481254260396e-05,
275
+ "loss": 1.9057,
276
+ "step": 380
277
+ },
278
+ {
279
+ "epoch": 0.2659506456974811,
280
+ "grad_norm": 16.25,
281
+ "learning_rate": 3.674164962508521e-05,
282
+ "loss": 2.3365,
283
+ "step": 390
284
+ },
285
+ {
286
+ "epoch": 0.2727698930230576,
287
+ "grad_norm": 8.625,
288
+ "learning_rate": 3.6400817995910025e-05,
289
+ "loss": 2.0162,
290
+ "step": 400
291
+ },
292
+ {
293
+ "epoch": 0.279589140348634,
294
+ "grad_norm": 13.5,
295
+ "learning_rate": 3.605998636673483e-05,
296
+ "loss": 2.1568,
297
+ "step": 410
298
+ },
299
+ {
300
+ "epoch": 0.28640838767421045,
301
+ "grad_norm": 11.3125,
302
+ "learning_rate": 3.571915473755965e-05,
303
+ "loss": 2.0719,
304
+ "step": 420
305
+ },
306
+ {
307
+ "epoch": 0.2932276349997869,
308
+ "grad_norm": 9.75,
309
+ "learning_rate": 3.537832310838446e-05,
310
+ "loss": 1.9952,
311
+ "step": 430
312
+ },
313
+ {
314
+ "epoch": 0.30004688232536336,
315
+ "grad_norm": 9.875,
316
+ "learning_rate": 3.503749147920927e-05,
317
+ "loss": 1.889,
318
+ "step": 440
319
+ },
320
+ {
321
+ "epoch": 0.3068661296509398,
322
+ "grad_norm": 10.5,
323
+ "learning_rate": 3.4696659850034084e-05,
324
+ "loss": 2.0325,
325
+ "step": 450
326
+ },
327
+ {
328
+ "epoch": 0.3136853769765162,
329
+ "grad_norm": 13.0,
330
+ "learning_rate": 3.43558282208589e-05,
331
+ "loss": 2.1277,
332
+ "step": 460
333
+ },
334
+ {
335
+ "epoch": 0.32050462430209264,
336
+ "grad_norm": 10.0,
337
+ "learning_rate": 3.401499659168371e-05,
338
+ "loss": 2.0542,
339
+ "step": 470
340
+ },
341
+ {
342
+ "epoch": 0.32732387162766907,
343
+ "grad_norm": 13.375,
344
+ "learning_rate": 3.367416496250852e-05,
345
+ "loss": 1.9533,
346
+ "step": 480
347
+ },
348
+ {
349
+ "epoch": 0.33414311895324555,
350
+ "grad_norm": 8.8125,
351
+ "learning_rate": 3.3333333333333335e-05,
352
+ "loss": 1.7744,
353
+ "step": 490
354
+ },
355
+ {
356
+ "epoch": 0.340962366278822,
357
+ "grad_norm": 11.375,
358
+ "learning_rate": 3.299250170415815e-05,
359
+ "loss": 1.9968,
360
+ "step": 500
361
+ },
362
+ {
363
+ "epoch": 0.3477816136043984,
364
+ "grad_norm": 10.625,
365
+ "learning_rate": 3.265167007498296e-05,
366
+ "loss": 1.7389,
367
+ "step": 510
368
+ },
369
+ {
370
+ "epoch": 0.35460086092997484,
371
+ "grad_norm": 11.0625,
372
+ "learning_rate": 3.231083844580777e-05,
373
+ "loss": 1.7532,
374
+ "step": 520
375
+ },
376
+ {
377
+ "epoch": 0.3614201082555513,
378
+ "grad_norm": 11.0625,
379
+ "learning_rate": 3.1970006816632586e-05,
380
+ "loss": 1.8736,
381
+ "step": 530
382
+ },
383
+ {
384
+ "epoch": 0.36823935558112775,
385
+ "grad_norm": 12.25,
386
+ "learning_rate": 3.16291751874574e-05,
387
+ "loss": 1.7424,
388
+ "step": 540
389
+ },
390
+ {
391
+ "epoch": 0.3750586029067042,
392
+ "grad_norm": 11.3125,
393
+ "learning_rate": 3.1288343558282215e-05,
394
+ "loss": 1.9172,
395
+ "step": 550
396
+ },
397
+ {
398
+ "epoch": 0.3818778502322806,
399
+ "grad_norm": 10.125,
400
+ "learning_rate": 3.094751192910702e-05,
401
+ "loss": 1.9809,
402
+ "step": 560
403
+ },
404
+ {
405
+ "epoch": 0.38869709755785703,
406
+ "grad_norm": 11.6875,
407
+ "learning_rate": 3.060668029993183e-05,
408
+ "loss": 1.6957,
409
+ "step": 570
410
+ },
411
+ {
412
+ "epoch": 0.3955163448834335,
413
+ "grad_norm": 12.5,
414
+ "learning_rate": 3.026584867075665e-05,
415
+ "loss": 1.9704,
416
+ "step": 580
417
+ },
418
+ {
419
+ "epoch": 0.40233559220900994,
420
+ "grad_norm": 11.0625,
421
+ "learning_rate": 2.992501704158146e-05,
422
+ "loss": 1.9209,
423
+ "step": 590
424
+ },
425
+ {
426
+ "epoch": 0.40915483953458637,
427
+ "grad_norm": 14.875,
428
+ "learning_rate": 2.9584185412406274e-05,
429
+ "loss": 1.8326,
430
+ "step": 600
431
+ },
432
+ {
433
+ "epoch": 0.4159740868601628,
434
+ "grad_norm": 12.6875,
435
+ "learning_rate": 2.9243353783231085e-05,
436
+ "loss": 1.8667,
437
+ "step": 610
438
+ },
439
+ {
440
+ "epoch": 0.4227933341857392,
441
+ "grad_norm": 14.3125,
442
+ "learning_rate": 2.89025221540559e-05,
443
+ "loss": 2.0266,
444
+ "step": 620
445
+ },
446
+ {
447
+ "epoch": 0.4296125815113157,
448
+ "grad_norm": 9.0625,
449
+ "learning_rate": 2.8561690524880707e-05,
450
+ "loss": 1.8784,
451
+ "step": 630
452
+ },
453
+ {
454
+ "epoch": 0.43643182883689213,
455
+ "grad_norm": 9.125,
456
+ "learning_rate": 2.822085889570552e-05,
457
+ "loss": 1.782,
458
+ "step": 640
459
+ },
460
+ {
461
+ "epoch": 0.44325107616246856,
462
+ "grad_norm": 15.4375,
463
+ "learning_rate": 2.7880027266530333e-05,
464
+ "loss": 1.7838,
465
+ "step": 650
466
+ },
467
+ {
468
+ "epoch": 0.450070323488045,
469
+ "grad_norm": 9.625,
470
+ "learning_rate": 2.7539195637355147e-05,
471
+ "loss": 1.8668,
472
+ "step": 660
473
+ },
474
+ {
475
+ "epoch": 0.45688957081362147,
476
+ "grad_norm": 16.75,
477
+ "learning_rate": 2.719836400817996e-05,
478
+ "loss": 1.9149,
479
+ "step": 670
480
+ },
481
+ {
482
+ "epoch": 0.4637088181391979,
483
+ "grad_norm": 11.0,
484
+ "learning_rate": 2.6857532379004773e-05,
485
+ "loss": 1.6853,
486
+ "step": 680
487
+ },
488
+ {
489
+ "epoch": 0.4705280654647743,
490
+ "grad_norm": 10.8125,
491
+ "learning_rate": 2.6516700749829587e-05,
492
+ "loss": 1.7238,
493
+ "step": 690
494
+ },
495
+ {
496
+ "epoch": 0.47734731279035075,
497
+ "grad_norm": 12.625,
498
+ "learning_rate": 2.6175869120654402e-05,
499
+ "loss": 1.7802,
500
+ "step": 700
501
+ },
502
+ {
503
+ "epoch": 0.4841665601159272,
504
+ "grad_norm": 9.1875,
505
+ "learning_rate": 2.583503749147921e-05,
506
+ "loss": 1.7323,
507
+ "step": 710
508
+ },
509
+ {
510
+ "epoch": 0.49098580744150366,
511
+ "grad_norm": 8.9375,
512
+ "learning_rate": 2.549420586230402e-05,
513
+ "loss": 1.7347,
514
+ "step": 720
515
+ },
516
+ {
517
+ "epoch": 0.4978050547670801,
518
+ "grad_norm": 14.5,
519
+ "learning_rate": 2.5153374233128835e-05,
520
+ "loss": 1.9333,
521
+ "step": 730
522
+ },
523
+ {
524
+ "epoch": 0.5046243020926565,
525
+ "grad_norm": 11.8125,
526
+ "learning_rate": 2.481254260395365e-05,
527
+ "loss": 1.799,
528
+ "step": 740
529
+ },
530
+ {
531
+ "epoch": 0.511443549418233,
532
+ "grad_norm": 12.1875,
533
+ "learning_rate": 2.447171097477846e-05,
534
+ "loss": 1.5268,
535
+ "step": 750
536
+ },
537
+ {
538
+ "epoch": 0.5182627967438094,
539
+ "grad_norm": 14.875,
540
+ "learning_rate": 2.4130879345603275e-05,
541
+ "loss": 1.4726,
542
+ "step": 760
543
+ },
544
+ {
545
+ "epoch": 0.5250820440693859,
546
+ "grad_norm": 9.1875,
547
+ "learning_rate": 2.3790047716428086e-05,
548
+ "loss": 1.6679,
549
+ "step": 770
550
+ },
551
+ {
552
+ "epoch": 0.5319012913949622,
553
+ "grad_norm": 11.75,
554
+ "learning_rate": 2.3449216087252897e-05,
555
+ "loss": 1.6178,
556
+ "step": 780
557
+ },
558
+ {
559
+ "epoch": 0.5387205387205387,
560
+ "grad_norm": 11.5625,
561
+ "learning_rate": 2.3108384458077712e-05,
562
+ "loss": 1.6491,
563
+ "step": 790
564
+ },
565
+ {
566
+ "epoch": 0.5455397860461152,
567
+ "grad_norm": 9.125,
568
+ "learning_rate": 2.2767552828902523e-05,
569
+ "loss": 1.6251,
570
+ "step": 800
571
+ },
572
+ {
573
+ "epoch": 0.5523590333716916,
574
+ "grad_norm": 10.875,
575
+ "learning_rate": 2.2426721199727334e-05,
576
+ "loss": 1.7639,
577
+ "step": 810
578
+ },
579
+ {
580
+ "epoch": 0.559178280697268,
581
+ "grad_norm": 12.875,
582
+ "learning_rate": 2.208588957055215e-05,
583
+ "loss": 1.5831,
584
+ "step": 820
585
+ },
586
+ {
587
+ "epoch": 0.5659975280228445,
588
+ "grad_norm": 10.1875,
589
+ "learning_rate": 2.1745057941376963e-05,
590
+ "loss": 1.6586,
591
+ "step": 830
592
+ },
593
+ {
594
+ "epoch": 0.5728167753484209,
595
+ "grad_norm": 11.9375,
596
+ "learning_rate": 2.140422631220177e-05,
597
+ "loss": 1.6956,
598
+ "step": 840
599
+ },
600
+ {
601
+ "epoch": 0.5796360226739974,
602
+ "grad_norm": 8.3125,
603
+ "learning_rate": 2.1063394683026585e-05,
604
+ "loss": 1.6436,
605
+ "step": 850
606
+ },
607
+ {
608
+ "epoch": 0.5864552699995738,
609
+ "grad_norm": 10.125,
610
+ "learning_rate": 2.07225630538514e-05,
611
+ "loss": 1.5014,
612
+ "step": 860
613
+ },
614
+ {
615
+ "epoch": 0.5932745173251502,
616
+ "grad_norm": 14.0625,
617
+ "learning_rate": 2.038173142467621e-05,
618
+ "loss": 1.5563,
619
+ "step": 870
620
+ },
621
+ {
622
+ "epoch": 0.6000937646507267,
623
+ "grad_norm": 12.25,
624
+ "learning_rate": 2.0040899795501022e-05,
625
+ "loss": 1.5834,
626
+ "step": 880
627
+ },
628
+ {
629
+ "epoch": 0.6069130119763031,
630
+ "grad_norm": 11.875,
631
+ "learning_rate": 1.9700068166325836e-05,
632
+ "loss": 1.6049,
633
+ "step": 890
634
+ },
635
+ {
636
+ "epoch": 0.6137322593018796,
637
+ "grad_norm": 10.375,
638
+ "learning_rate": 1.935923653715065e-05,
639
+ "loss": 1.5865,
640
+ "step": 900
641
+ },
642
+ {
643
+ "epoch": 0.620551506627456,
644
+ "grad_norm": 12.0625,
645
+ "learning_rate": 1.9018404907975462e-05,
646
+ "loss": 1.5695,
647
+ "step": 910
648
+ },
649
+ {
650
+ "epoch": 0.6273707539530324,
651
+ "grad_norm": 10.6875,
652
+ "learning_rate": 1.8677573278800273e-05,
653
+ "loss": 1.484,
654
+ "step": 920
655
+ },
656
+ {
657
+ "epoch": 0.6341900012786089,
658
+ "grad_norm": 12.0,
659
+ "learning_rate": 1.8336741649625087e-05,
660
+ "loss": 1.4748,
661
+ "step": 930
662
+ },
663
+ {
664
+ "epoch": 0.6410092486041853,
665
+ "grad_norm": 9.625,
666
+ "learning_rate": 1.79959100204499e-05,
667
+ "loss": 1.4673,
668
+ "step": 940
669
+ },
670
+ {
671
+ "epoch": 0.6478284959297618,
672
+ "grad_norm": 12.3125,
673
+ "learning_rate": 1.7655078391274713e-05,
674
+ "loss": 1.4539,
675
+ "step": 950
676
+ },
677
+ {
678
+ "epoch": 0.6546477432553381,
679
+ "grad_norm": 10.9375,
680
+ "learning_rate": 1.7314246762099524e-05,
681
+ "loss": 1.6066,
682
+ "step": 960
683
+ },
684
+ {
685
+ "epoch": 0.6614669905809146,
686
+ "grad_norm": 12.8125,
687
+ "learning_rate": 1.6973415132924335e-05,
688
+ "loss": 1.6063,
689
+ "step": 970
690
+ },
691
+ {
692
+ "epoch": 0.6682862379064911,
693
+ "grad_norm": 151.0,
694
+ "learning_rate": 1.663258350374915e-05,
695
+ "loss": 1.4019,
696
+ "step": 980
697
+ },
698
+ {
699
+ "epoch": 0.6751054852320675,
700
+ "grad_norm": 10.875,
701
+ "learning_rate": 1.629175187457396e-05,
702
+ "loss": 1.4869,
703
+ "step": 990
704
+ },
705
+ {
706
+ "epoch": 0.681924732557644,
707
+ "grad_norm": 11.375,
708
+ "learning_rate": 1.5950920245398772e-05,
709
+ "loss": 1.5375,
710
+ "step": 1000
711
+ },
712
+ {
713
+ "epoch": 0.6887439798832203,
714
+ "grad_norm": 13.3125,
715
+ "learning_rate": 1.5610088616223586e-05,
716
+ "loss": 1.4521,
717
+ "step": 1010
718
+ },
719
+ {
720
+ "epoch": 0.6955632272087968,
721
+ "grad_norm": 11.375,
722
+ "learning_rate": 1.52692569870484e-05,
723
+ "loss": 1.5836,
724
+ "step": 1020
725
+ },
726
+ {
727
+ "epoch": 0.7023824745343733,
728
+ "grad_norm": 8.6875,
729
+ "learning_rate": 1.492842535787321e-05,
730
+ "loss": 1.4212,
731
+ "step": 1030
732
+ },
733
+ {
734
+ "epoch": 0.7092017218599497,
735
+ "grad_norm": 10.125,
736
+ "learning_rate": 1.4587593728698023e-05,
737
+ "loss": 1.5075,
738
+ "step": 1040
739
+ },
740
+ {
741
+ "epoch": 0.7160209691855262,
742
+ "grad_norm": 12.4375,
743
+ "learning_rate": 1.4246762099522837e-05,
744
+ "loss": 1.3275,
745
+ "step": 1050
746
+ },
747
+ {
748
+ "epoch": 0.7228402165111026,
749
+ "grad_norm": 10.0625,
750
+ "learning_rate": 1.390593047034765e-05,
751
+ "loss": 1.278,
752
+ "step": 1060
753
+ },
754
+ {
755
+ "epoch": 0.729659463836679,
756
+ "grad_norm": 14.125,
757
+ "learning_rate": 1.3565098841172461e-05,
758
+ "loss": 1.2336,
759
+ "step": 1070
760
+ },
761
+ {
762
+ "epoch": 0.7364787111622555,
763
+ "grad_norm": 9.9375,
764
+ "learning_rate": 1.3224267211997274e-05,
765
+ "loss": 1.3917,
766
+ "step": 1080
767
+ },
768
+ {
769
+ "epoch": 0.7432979584878319,
770
+ "grad_norm": 10.25,
771
+ "learning_rate": 1.2883435582822087e-05,
772
+ "loss": 1.3885,
773
+ "step": 1090
774
+ },
775
+ {
776
+ "epoch": 0.7501172058134083,
777
+ "grad_norm": 9.9375,
778
+ "learning_rate": 1.25426039536469e-05,
779
+ "loss": 1.3438,
780
+ "step": 1100
781
+ },
782
+ {
783
+ "epoch": 0.7569364531389848,
784
+ "grad_norm": 10.1875,
785
+ "learning_rate": 1.2201772324471712e-05,
786
+ "loss": 1.4968,
787
+ "step": 1110
788
+ },
789
+ {
790
+ "epoch": 0.7637557004645612,
791
+ "grad_norm": 12.1875,
792
+ "learning_rate": 1.1860940695296524e-05,
793
+ "loss": 1.3124,
794
+ "step": 1120
795
+ },
796
+ {
797
+ "epoch": 0.7705749477901377,
798
+ "grad_norm": 9.8125,
799
+ "learning_rate": 1.1520109066121336e-05,
800
+ "loss": 1.282,
801
+ "step": 1130
802
+ },
803
+ {
804
+ "epoch": 0.7773941951157141,
805
+ "grad_norm": 12.8125,
806
+ "learning_rate": 1.1179277436946149e-05,
807
+ "loss": 1.2937,
808
+ "step": 1140
809
+ },
810
+ {
811
+ "epoch": 0.7842134424412905,
812
+ "grad_norm": 10.5,
813
+ "learning_rate": 1.0838445807770962e-05,
814
+ "loss": 1.3717,
815
+ "step": 1150
816
+ },
817
+ {
818
+ "epoch": 0.791032689766867,
819
+ "grad_norm": 10.5625,
820
+ "learning_rate": 1.0497614178595775e-05,
821
+ "loss": 1.2819,
822
+ "step": 1160
823
+ },
824
+ {
825
+ "epoch": 0.7978519370924434,
826
+ "grad_norm": 15.75,
827
+ "learning_rate": 1.0156782549420586e-05,
828
+ "loss": 1.3,
829
+ "step": 1170
830
+ },
831
+ {
832
+ "epoch": 0.8046711844180199,
833
+ "grad_norm": 9.0,
834
+ "learning_rate": 9.8159509202454e-06,
835
+ "loss": 1.4474,
836
+ "step": 1180
837
+ },
838
+ {
839
+ "epoch": 0.8114904317435963,
840
+ "grad_norm": 8.5,
841
+ "learning_rate": 9.475119291070211e-06,
842
+ "loss": 1.412,
843
+ "step": 1190
844
+ },
845
+ {
846
+ "epoch": 0.8183096790691727,
847
+ "grad_norm": 10.75,
848
+ "learning_rate": 9.134287661895024e-06,
849
+ "loss": 1.2744,
850
+ "step": 1200
851
+ },
852
+ {
853
+ "epoch": 0.8251289263947492,
854
+ "grad_norm": 8.8125,
855
+ "learning_rate": 8.793456032719837e-06,
856
+ "loss": 1.2013,
857
+ "step": 1210
858
+ },
859
+ {
860
+ "epoch": 0.8319481737203256,
861
+ "grad_norm": 17.125,
862
+ "learning_rate": 8.45262440354465e-06,
863
+ "loss": 1.463,
864
+ "step": 1220
865
+ },
866
+ {
867
+ "epoch": 0.8387674210459021,
868
+ "grad_norm": 11.125,
869
+ "learning_rate": 8.11179277436946e-06,
870
+ "loss": 1.3798,
871
+ "step": 1230
872
+ },
873
+ {
874
+ "epoch": 0.8455866683714784,
875
+ "grad_norm": 11.75,
876
+ "learning_rate": 7.770961145194275e-06,
877
+ "loss": 1.2087,
878
+ "step": 1240
879
+ },
880
+ {
881
+ "epoch": 0.8524059156970549,
882
+ "grad_norm": 10.125,
883
+ "learning_rate": 7.430129516019086e-06,
884
+ "loss": 1.3037,
885
+ "step": 1250
886
+ },
887
+ {
888
+ "epoch": 0.8592251630226314,
889
+ "grad_norm": 9.375,
890
+ "learning_rate": 7.0892978868439e-06,
891
+ "loss": 1.3177,
892
+ "step": 1260
893
+ },
894
+ {
895
+ "epoch": 0.8660444103482078,
896
+ "grad_norm": 14.0625,
897
+ "learning_rate": 6.748466257668712e-06,
898
+ "loss": 1.5506,
899
+ "step": 1270
900
+ },
901
+ {
902
+ "epoch": 0.8728636576737843,
903
+ "grad_norm": 9.5,
904
+ "learning_rate": 6.4076346284935256e-06,
905
+ "loss": 1.4374,
906
+ "step": 1280
907
+ },
908
+ {
909
+ "epoch": 0.8796829049993607,
910
+ "grad_norm": 14.125,
911
+ "learning_rate": 6.066802999318337e-06,
912
+ "loss": 1.2055,
913
+ "step": 1290
914
+ },
915
+ {
916
+ "epoch": 0.8865021523249371,
917
+ "grad_norm": 13.75,
918
+ "learning_rate": 5.7259713701431494e-06,
919
+ "loss": 1.4779,
920
+ "step": 1300
921
+ },
922
+ {
923
+ "epoch": 0.8933213996505136,
924
+ "grad_norm": 12.1875,
925
+ "learning_rate": 5.385139740967962e-06,
926
+ "loss": 1.3483,
927
+ "step": 1310
928
+ },
929
+ {
930
+ "epoch": 0.90014064697609,
931
+ "grad_norm": 12.4375,
932
+ "learning_rate": 5.044308111792775e-06,
933
+ "loss": 1.3771,
934
+ "step": 1320
935
+ },
936
+ {
937
+ "epoch": 0.9069598943016665,
938
+ "grad_norm": 9.3125,
939
+ "learning_rate": 4.703476482617587e-06,
940
+ "loss": 1.5074,
941
+ "step": 1330
942
+ },
943
+ {
944
+ "epoch": 0.9137791416272429,
945
+ "grad_norm": 9.8125,
946
+ "learning_rate": 4.3626448534424e-06,
947
+ "loss": 1.1781,
948
+ "step": 1340
949
+ },
950
+ {
951
+ "epoch": 0.9205983889528193,
952
+ "grad_norm": 9.4375,
953
+ "learning_rate": 4.0218132242672125e-06,
954
+ "loss": 1.3096,
955
+ "step": 1350
956
+ },
957
+ {
958
+ "epoch": 0.9274176362783958,
959
+ "grad_norm": 10.0,
960
+ "learning_rate": 3.680981595092025e-06,
961
+ "loss": 1.4104,
962
+ "step": 1360
963
+ },
964
+ {
965
+ "epoch": 0.9342368836039722,
966
+ "grad_norm": 10.8125,
967
+ "learning_rate": 3.3401499659168377e-06,
968
+ "loss": 1.3616,
969
+ "step": 1370
970
+ },
971
+ {
972
+ "epoch": 0.9410561309295487,
973
+ "grad_norm": 12.4375,
974
+ "learning_rate": 2.9993183367416496e-06,
975
+ "loss": 1.2745,
976
+ "step": 1380
977
+ },
978
+ {
979
+ "epoch": 0.9478753782551251,
980
+ "grad_norm": 11.5,
981
+ "learning_rate": 2.6584867075664624e-06,
982
+ "loss": 1.1573,
983
+ "step": 1390
984
+ },
985
+ {
986
+ "epoch": 0.9546946255807015,
987
+ "grad_norm": 11.1875,
988
+ "learning_rate": 2.3176550783912747e-06,
989
+ "loss": 1.2449,
990
+ "step": 1400
991
+ },
992
+ {
993
+ "epoch": 0.961513872906278,
994
+ "grad_norm": 9.9375,
995
+ "learning_rate": 1.9768234492160875e-06,
996
+ "loss": 1.2238,
997
+ "step": 1410
998
+ },
999
+ {
1000
+ "epoch": 0.9683331202318544,
1001
+ "grad_norm": 9.0,
1002
+ "learning_rate": 1.6359918200409001e-06,
1003
+ "loss": 1.2504,
1004
+ "step": 1420
1005
+ },
1006
+ {
1007
+ "epoch": 0.9751523675574308,
1008
+ "grad_norm": 9.5,
1009
+ "learning_rate": 1.2951601908657125e-06,
1010
+ "loss": 1.253,
1011
+ "step": 1430
1012
+ },
1013
+ {
1014
+ "epoch": 0.9819716148830073,
1015
+ "grad_norm": 8.75,
1016
+ "learning_rate": 9.54328561690525e-07,
1017
+ "loss": 1.2834,
1018
+ "step": 1440
1019
+ },
1020
+ {
1021
+ "epoch": 0.9887908622085837,
1022
+ "grad_norm": 6.9375,
1023
+ "learning_rate": 6.134969325153375e-07,
1024
+ "loss": 1.2181,
1025
+ "step": 1450
1026
+ },
1027
+ {
1028
+ "epoch": 0.9956101095341602,
1029
+ "grad_norm": 9.6875,
1030
+ "learning_rate": 2.7266530334015e-07,
1031
+ "loss": 1.3088,
1032
+ "step": 1460
1033
+ }
1034
+ ],
1035
+ "logging_steps": 10,
1036
+ "max_steps": 1467,
1037
+ "num_input_tokens_seen": 0,
1038
+ "num_train_epochs": 1,
1039
+ "save_steps": 100,
1040
+ "stateful_callbacks": {
1041
+ "TrainerControl": {
1042
+ "args": {
1043
+ "should_epoch_stop": false,
1044
+ "should_evaluate": false,
1045
+ "should_log": false,
1046
+ "should_save": true,
1047
+ "should_training_stop": true
1048
+ },
1049
+ "attributes": {}
1050
+ }
1051
+ },
1052
+ "total_flos": 3.6373456607845786e+17,
1053
+ "train_batch_size": 1,
1054
+ "trial_name": null,
1055
+ "trial_params": null
1056
+ }
checkpoint-1467/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbe1c1a78d462e3303123f4ff7facf76d33471dad0aab240d4af4d3300587940
3
+ size 5432
config.json ADDED
@@ -0,0 +1,148 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_attn_implementation_autoset": false,
3
+ "acoustic_tokenizer_config": {
4
+ "causal": true,
5
+ "channels": 1,
6
+ "conv_bias": true,
7
+ "conv_norm": "none",
8
+ "corpus_normalize": 0.0,
9
+ "decoder_depths": null,
10
+ "decoder_n_filters": 32,
11
+ "decoder_ratios": [
12
+ 8,
13
+ 5,
14
+ 5,
15
+ 4,
16
+ 2,
17
+ 2
18
+ ],
19
+ "disable_last_norm": true,
20
+ "dtype": "bfloat16",
21
+ "encoder_depths": "3-3-3-3-3-3-8",
22
+ "encoder_n_filters": 32,
23
+ "encoder_ratios": [
24
+ 8,
25
+ 5,
26
+ 5,
27
+ 4,
28
+ 2,
29
+ 2
30
+ ],
31
+ "fix_std": 0.5,
32
+ "layer_scale_init_value": 1e-06,
33
+ "layernorm": "RMSNorm",
34
+ "layernorm_elementwise_affine": true,
35
+ "layernorm_eps": 1e-05,
36
+ "mixer_layer": "depthwise_conv",
37
+ "model_type": "vibevoice_acoustic_tokenizer",
38
+ "pad_mode": "constant",
39
+ "std_dist_type": "gaussian",
40
+ "vae_dim": 64,
41
+ "weight_init_value": 0.01
42
+ },
43
+ "acoustic_vae_dim": 64,
44
+ "architectures": [
45
+ "VibeVoiceASRForConditionalGeneration"
46
+ ],
47
+ "decoder_config": {
48
+ "attention_dropout": 0.0,
49
+ "dtype": "bfloat16",
50
+ "hidden_act": "silu",
51
+ "hidden_size": 3584,
52
+ "initializer_range": 0.02,
53
+ "intermediate_size": 18944,
54
+ "layer_types": [
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention",
72
+ "full_attention",
73
+ "full_attention",
74
+ "full_attention",
75
+ "full_attention",
76
+ "full_attention",
77
+ "full_attention",
78
+ "full_attention",
79
+ "full_attention",
80
+ "full_attention",
81
+ "full_attention",
82
+ "full_attention"
83
+ ],
84
+ "max_position_embeddings": 131072,
85
+ "max_window_layers": 28,
86
+ "model_type": "qwen2",
87
+ "num_attention_heads": 28,
88
+ "num_hidden_layers": 28,
89
+ "num_key_value_heads": 4,
90
+ "rms_norm_eps": 1e-06,
91
+ "rope_scaling": null,
92
+ "rope_theta": 1000000.0,
93
+ "sliding_window": null,
94
+ "use_cache": true,
95
+ "use_mrope": false,
96
+ "use_sliding_window": false,
97
+ "vocab_size": 152064
98
+ },
99
+ "diffusion_head_config": {
100
+ "ddpm_batch_mul": 4,
101
+ "ddpm_beta_schedule": "cosine",
102
+ "ddpm_num_inference_steps": 20,
103
+ "ddpm_num_steps": 1000,
104
+ "diffusion_type": "ddpm",
105
+ "head_ffn_ratio": 3.0,
106
+ "head_layers": 4,
107
+ "hidden_size": 3584,
108
+ "latent_size": 64,
109
+ "model_type": "vibepod_diffusion_head",
110
+ "prediction_type": "v_prediction",
111
+ "rms_norm_eps": 1e-05,
112
+ "speech_vae_dim": 64
113
+ },
114
+ "dtype": "bfloat16",
115
+ "model_type": "vibevoice",
116
+ "semantic_tokenizer_config": {
117
+ "causal": true,
118
+ "channels": 1,
119
+ "conv_bias": true,
120
+ "conv_norm": "none",
121
+ "corpus_normalize": 0.0,
122
+ "disable_last_norm": true,
123
+ "dtype": "bfloat16",
124
+ "encoder_depths": "3-3-3-3-3-3-8",
125
+ "encoder_n_filters": 32,
126
+ "encoder_ratios": [
127
+ 8,
128
+ 5,
129
+ 5,
130
+ 4,
131
+ 2,
132
+ 2
133
+ ],
134
+ "fix_std": 0,
135
+ "layer_scale_init_value": 1e-06,
136
+ "layernorm": "RMSNorm",
137
+ "layernorm_elementwise_affine": true,
138
+ "layernorm_eps": 1e-05,
139
+ "mixer_layer": "depthwise_conv",
140
+ "model_type": "vibevoice_semantic_tokenizer",
141
+ "pad_mode": "constant",
142
+ "std_dist_type": "none",
143
+ "vae_dim": 128,
144
+ "weight_init_value": 0.01
145
+ },
146
+ "semantic_vae_dim": 128,
147
+ "transformers_version": "4.56.2"
148
+ }
generation_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "transformers_version": "4.56.2"
4
+ }
job_metadata.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_id": "microsoft/VibeVoice-ASR",
3
+ "dataset_id": "laion/laions_got_talent_gpt4o_voice_acting_validated_vocal_bursts",
4
+ "language_code": "en",
5
+ "max_audio_length": 35.0,
6
+ "num_train_epochs": 1,
7
+ "per_device_train_batch_size": 1,
8
+ "gradient_accumulation_steps": 16,
9
+ "learning_rate": 5e-05,
10
+ "limit": 0,
11
+ "canonicalize_events": true,
12
+ "family_repeat_map": "laugh:1,sigh:4,breath:2,cough:3,moan:1,cry:2",
13
+ "family_cap_map": "laugh:9000,breath:7000,moan:5000,sigh:7000,cough:5000,cry:5000",
14
+ "training_mode": "full_finetune"
15
+ }
model-00001-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c70cfe88d16fbd345fe44bdf879c797a7b6ec50e94a024f900cb867adc48e0f
3
+ size 4877662336
model-00002-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ee1986d89ea278d2f19e847cf938595ae2590764e199b7dfcd6baced0fd3002
3
+ size 4932752840
model-00003-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1733708cf7f6773d071fe61a8364c3ab5cb9a27a9d22fb4ad67511ba061a2d6
3
+ size 4982901128
model-00004-of-00004.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9649879a4c80f30f7bc0509a9662c92c626016f2565d4e91aeb1dbe88287f9bd
3
+ size 2554883402
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
preprocessor_config.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "processor_class": "VibeVoiceASRProcessor",
3
+ "speech_tok_compress_ratio": 3200,
4
+ "target_sample_rate": 24000,
5
+ "normalize_audio": true,
6
+ "target_dB_FS": -25,
7
+ "eps": 1e-06
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 1.0,
3
+ "total_flos": 3.6373456607845786e+17,
4
+ "train_loss": 1.8379220435955772,
5
+ "train_runtime": 9787.7529,
6
+ "train_samples_per_second": 2.397,
7
+ "train_steps_per_second": 0.15
8
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbe1c1a78d462e3303123f4ff7facf76d33471dad0aab240d4af4d3300587940
3
+ size 5432