Hugging Face's logo

mrjackspade
/
Ideogram4-Natural-Language-Text-Encoder

Text-to-Image
comfyui
ideogram4
qwen3-vl
text-encoder
lora
fp8
flow-matching
experimental
Model card Files
xet
 Community
mrjackspade commited on
Commit
546e1b0
·
verified ·
1 Parent(s): 2fda0d1

Publish step 1530 checkpoint manifest

Browse files
Files changed (1) hide show
  1. manifests/checkpoint_step_00001530.json +156 -0
manifests/checkpoint_step_00001530.json ADDED
@@ -0,0 +1,156 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "status": "complete",
4
+ "config_fingerprint": "0667d4ea7a9eeb08101909c8f4b52860faadd491ce6d4ceb8094405e04a5f75e",
5
+ "dataset_fingerprint": "6cf9eb3d777a338eb007dcc5606b2aac6605bf899a191ea399ccd77d5527acfb",
6
+ "cursor": {
7
+ "global_step": 1530,
8
+ "micro_step": 12275,
9
+ "dataset_epoch": 0,
10
+ "sample_offset": 24480,
11
+ "accumulation_step": 0
12
+ },
13
+ "extra": {
14
+ "last_metrics": {
15
+ "global_step": 1530,
16
+ "behavior_loss": 0.01784470462007448,
17
+ "behavior_microbatches": 8,
18
+ "behavior_singleton_microbatches": 0,
19
+ "oom_replayed_as_singletons": false,
20
+ "gradient_norm": 0.017304301261901855,
21
+ "learning_rate_used": 5e-06,
22
+ "step_seconds": 12.875,
23
+ "peak_reserved_gib": 24.904,
24
+ "optimizer_state_device": "cpu_between_updates",
25
+ "weighted_behavior_objective": 0.01784470462007448,
26
+ "validation_behavior_loss": 0.017060593152418733,
27
+ "lr_window_phase": "confirming",
28
+ "lr_window_start_step": 1480,
29
+ "lr_window_end_step": 1530,
30
+ "lr_window_point_count": 6,
31
+ "lr_window_log_slope_per_step": 2.7408283244737912e-05,
32
+ "lr_window_fitted_log_descent": -0.0013704141622368955,
33
+ "lr_window_relative_descent": -0.00137135360881957,
34
+ "lr_window_residual_mad_scale": 0.0010623870441984341,
35
+ "lr_window_descent_to_noise": -1.2899386995733428,
36
+ "lr_window_accepted": false,
37
+ "lr_window_elbow_step": 1490,
38
+ "lr_training_window_start_step": 1490,
39
+ "lr_training_window_end_step": 1530,
40
+ "lr_training_window_point_count": 5,
41
+ "lr_training_window_log_slope_per_step": 0.003067318613296203,
42
+ "lr_training_window_fitted_log_descent": -0.12269274453184813,
43
+ "lr_training_window_relative_descent": -0.13053700390093237,
44
+ "lr_training_window_residual_mad_scale": 0.0651870427677517,
45
+ "lr_training_window_descent_to_noise": -1.8821646039225566,
46
+ "lr_training_window_accepted": false,
47
+ "lr_training_window_elbow_step": 1500,
48
+ "lr_window_any_signal_accepted": false,
49
+ "validation_interval": 10,
50
+ "next_validation_step": 1540,
51
+ "learning_rate": 5e-06
52
+ },
53
+ "reason": "periodic",
54
+ "parent": null,
55
+ "lr_control": {
56
+ "version": 3,
57
+ "config": {
58
+ "factor": 0.5,
59
+ "minimum_learning_rate": 1e-06,
60
+ "window_steps": 50,
61
+ "probe_every_steps": 10,
62
+ "confirmation_steps": 20,
63
+ "minimum_descent_to_noise": 1.0,
64
+ "minimum_relative_descent": 0.0
65
+ },
66
+ "state": {
67
+ "learning_rate": 5e-06,
68
+ "points": [
69
+ {
70
+ "step": 1480,
71
+ "loss": 0.0170914699556306
72
+ },
73
+ {
74
+ "step": 1490,
75
+ "loss": 0.017028517997823657
76
+ },
77
+ {
78
+ "step": 1500,
79
+ "loss": 0.017046570871025325
80
+ },
81
+ {
82
+ "step": 1510,
83
+ "loss": 0.017040204368531705
84
+ },
85
+ {
86
+ "step": 1520,
87
+ "loss": 0.017080724351108075
88
+ },
89
+ {
90
+ "step": 1530,
91
+ "loss": 0.017060593152418733
92
+ }
93
+ ],
94
+ "training_points": [
95
+ {
96
+ "step": 1490,
97
+ "loss": 0.01357189457048662
98
+ },
99
+ {
100
+ "step": 1500,
101
+ "loss": 0.013392652114271186
102
+ },
103
+ {
104
+ "step": 1510,
105
+ "loss": 0.014630921845673583
106
+ },
107
+ {
108
+ "step": 1520,
109
+ "loss": 0.01602509914082475
110
+ },
111
+ {
112
+ "step": 1530,
113
+ "loss": 0.014382915897294879
114
+ }
115
+ ],
116
+ "training_loss_sum": 0.0,
117
+ "training_loss_count": 0,
118
+ "last_training_step": 1530,
119
+ "window_start_step": 1480,
120
+ "next_validation_step": 1540,
121
+ "last_analysis": {
122
+ "start_step": 1480,
123
+ "end_step": 1530,
124
+ "point_count": 6,
125
+ "log_slope_per_step": 2.7408283244737912e-05,
126
+ "fitted_log_descent": -0.0013704141622368955,
127
+ "relative_descent": -0.00137135360881957,
128
+ "residual_mad_scale": 0.0010623870441984341,
129
+ "descent_to_noise": -1.2899386995733428,
130
+ "accepted": false,
131
+ "elbow_step": 1490
132
+ },
133
+ "last_training_analysis": {
134
+ "start_step": 1490,
135
+ "end_step": 1530,
136
+ "point_count": 5,
137
+ "log_slope_per_step": 0.003067318613296203,
138
+ "fitted_log_descent": -0.12269274453184813,
139
+ "relative_descent": -0.13053700390093237,
140
+ "residual_mad_scale": 0.0651870427677517,
141
+ "descent_to_noise": -1.8821646039225566,
142
+ "accepted": false,
143
+ "elbow_step": 1500
144
+ },
145
+ "confirming": true,
146
+ "pending_rollback": null,
147
+ "reductions": 6
148
+ }
149
+ },
150
+ "rollback_replay": null
151
+ },
152
+ "files": {
153
+ "adapters.safetensors": "eb53557a4d1476bf2b62ac6946476bc4fcbf8282034ab04df1e6a33feceaa96c",
154
+ "training_state.pt": "1fc980b573bfc358cc596786e8c5162a4d6cb44a4dfdd945923cf498cd58b447"
155
+ }
156
+ }