Hugging Face's logo

explcre
/
ntv3_ar_650m_unconditional_t1

dna
enhancer-generation
mdlm
autoregressive
ntv3
unconditional-baseline
Model card Files
xet
 Community
explcre commited on
Commit
9bae83d
·
verified ·
1 Parent(s): 9ad5645

Add files using upload-large-folder tool

Browse files
Files changed (6) hide show
  1. README.md +54 -0
  2. best.pt +3 -0
  3. manifest.json +27 -0
  4. metrics.json +212 -0
  5. predictions_preview.jsonl +0 -0
  6. training_log.jsonl +422 -0
README.md ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ tags:
4
+ - dna
5
+ - enhancer-generation
6
+ - mdlm
7
+ - autoregressive
8
+ - ntv3
9
+ - unconditional-baseline
10
+ ---
11
+
12
+ # NTV3-AR-650M unconditional enhancer generator (T1)
13
+
14
+ Unconditional NTv3 generation baseline for the brain 7-cell
15
+ enhancer task. There is **no cell-type signal** at training or
16
+ inference time, so this model learns ``p(enhancer)`` and emits a
17
+ single mode-collapsed enhancer for every gold input regardless of
18
+ the requested cell type. Useful as a lower bound for cell-type
19
+ specificity (see the MDLM/AR rows in
20
+ [explcre/biomodel_reasoning_calling_study2](https://github.com/explcre/biomodel_reasoning_calling_study2)
21
+ ``regureasoner_loop/docs/EXPERIMENTS.md``).
22
+
23
+ Trained on `train.enhancer_generation.strat7c.n35k` (35k rows
24
+ stratified across 7 brain cell types). MDLM-8m / MDLM-650m use the
25
+ weighted-CE diffusion loss (1/t schedule); AR-650m uses standard
26
+ next-token CE.
27
+
28
+ ## Files
29
+
30
+ ```
31
+ best.pt # EMA / best-checkpoint state dict
32
+ metrics.json # 28-column eval against gold n7k held-out test
33
+ training_log.jsonl # per-step train metrics
34
+ manifest.json # run config snapshot
35
+ predictions_preview.jsonl # first 200 sampled rows (for spot-check)
36
+ ```
37
+
38
+ ## Headline eval (held-out gold n7k)
39
+
40
+ | parse_rate | argmax_acc | on_target | off_target | specificity | FID |
41
+ |---:|---:|---:|---:|---:|---:|
42
+ | 1.000 | 0.150 | -16.296 | -16.339 | 0.042 | 75.584 |
43
+
44
+
45
+ Random argmax baseline over 7 cells = 0.143; the unconditional
46
+ baselines all sit at chance because the model has no signal to
47
+ differentiate cell types. FID is misleadingly low (the average
48
+ enhancer is statistically plausible) but cell-type recall is
49
+ absent — see the cell-type-conditioned MDLM repos for the
50
+ conditional fix.
51
+
52
+ ## Citation
53
+ See the parent repo's `EXPERIMENTS.md` for the full per-baseline
54
+ comparison table.
best.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2ebc17bf280e9d45936a9bdebe40881ddd7e214fb2cc9d0d323bca82ade2f75f
3
+ size 1317642957
manifest.json ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_name": "train_loss",
3
+ "mode": "min",
4
+ "records": [
5
+ {
6
+ "path": "/extra/zhanglab0/INDV/pengchx3/regureasoner_loop/runs/gen_ar_enhancer_generation_20260429_130402/ckpt_ar_step001800_train_loss0.0005.pt",
7
+ "step": 1800,
8
+ "metric": 0.0004560278612188995,
9
+ "metric_name": "train_loss",
10
+ "extra": {}
11
+ },
12
+ {
13
+ "path": "/extra/zhanglab0/INDV/pengchx3/regureasoner_loop/runs/gen_ar_enhancer_generation_20260429_130402/ckpt_ar_step001600_train_loss0.0006.pt",
14
+ "step": 1600,
15
+ "metric": 0.0006041897577233613,
16
+ "metric_name": "train_loss",
17
+ "extra": {}
18
+ },
19
+ {
20
+ "path": "/extra/zhanglab0/INDV/pengchx3/regureasoner_loop/runs/gen_ar_enhancer_generation_20260429_130402/ckpt_ar_step001400_train_loss0.0010.pt",
21
+ "step": 1400,
22
+ "metric": 0.0009589027031324804,
23
+ "metric_name": "train_loss",
24
+ "extra": {}
25
+ }
26
+ ]
27
+ }
metrics.json ADDED
@@ -0,0 +1,212 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "basic_overall": {
3
+ "n": 7000,
4
+ "parse_rate": 1.0,
5
+ "mean_pred_len": 1024.0,
6
+ "mean_pred_gc": 0.5400828683035714,
7
+ "mean_gold_len": 500.0,
8
+ "mean_gold_gc": 0.5223668571428571,
9
+ "mean_gc_abs_err": 0.08570922098214286,
10
+ "mean_length_ratio": 2.048000000000001
11
+ },
12
+ "basic_per_celltype": {
13
+ "Ex": {
14
+ "n": 1000,
15
+ "parse_rate": 1.0,
16
+ "mean_pred_len": 1024.0,
17
+ "mean_pred_gc": 0.5391396484375
18
+ },
19
+ "In": {
20
+ "n": 1000,
21
+ "parse_rate": 1.0,
22
+ "mean_pred_len": 1024.0,
23
+ "mean_pred_gc": 0.54087890625
24
+ },
25
+ "OPC": {
26
+ "n": 1000,
27
+ "parse_rate": 1.0,
28
+ "mean_pred_len": 1024.0,
29
+ "mean_pred_gc": 0.5410673828125
30
+ },
31
+ "Ast": {
32
+ "n": 1000,
33
+ "parse_rate": 1.0,
34
+ "mean_pred_len": 1024.0,
35
+ "mean_pred_gc": 0.540087890625
36
+ },
37
+ "Oli": {
38
+ "n": 1000,
39
+ "parse_rate": 1.0,
40
+ "mean_pred_len": 1024.0,
41
+ "mean_pred_gc": 0.5399501953125
42
+ },
43
+ "Mic": {
44
+ "n": 1000,
45
+ "parse_rate": 1.0,
46
+ "mean_pred_len": 1024.0,
47
+ "mean_pred_gc": 0.5399033203125
48
+ },
49
+ "End": {
50
+ "n": 1000,
51
+ "parse_rate": 1.0,
52
+ "mean_pred_len": 1024.0,
53
+ "mean_pred_gc": 0.539552734375
54
+ }
55
+ },
56
+ "sequence_validity": {
57
+ "mean_length": 1024.0,
58
+ "max_length": 1024.0,
59
+ "min_length": 1024.0,
60
+ "mean_fraction_canonical": 1.0,
61
+ "max_fraction_canonical": 1.0,
62
+ "min_fraction_canonical": 1.0,
63
+ "mean_gc": 0.5400828683035714,
64
+ "max_gc": 0.6005859375,
65
+ "min_gc": 0.4892578125,
66
+ "mean_entropy_bits": 1.3146158034232096,
67
+ "max_entropy_bits": 1.3919739089107828,
68
+ "min_entropy_bits": 1.2370558577498938,
69
+ "mean_longest_homopolymer": 10.053571428571429,
70
+ "max_longest_homopolymer": 23.0,
71
+ "min_longest_homopolymer": 6.0,
72
+ "mean_tandem_repeat_fraction": 0.1808775111607143,
73
+ "max_tandem_repeat_fraction": 0.2958984375,
74
+ "min_tandem_repeat_fraction": 0.080078125
75
+ },
76
+ "sequence_validity_gold_reference": {
77
+ "mean_length": 500.0,
78
+ "max_length": 500.0,
79
+ "min_length": 500.0,
80
+ "mean_fraction_canonical": 1.0,
81
+ "max_fraction_canonical": 1.0,
82
+ "min_fraction_canonical": 1.0,
83
+ "mean_gc": 0.5223668571428571,
84
+ "max_gc": 0.896,
85
+ "min_gc": 0.276,
86
+ "mean_entropy_bits": 1.954767692123513,
87
+ "max_entropy_bits": 1.9999769166331194,
88
+ "min_entropy_bits": 1.4728614433169502,
89
+ "mean_longest_homopolymer": 7.143714285714286,
90
+ "max_longest_homopolymer": 42.0,
91
+ "min_longest_homopolymer": 3.0,
92
+ "mean_tandem_repeat_fraction": 0.04596485714285715,
93
+ "max_tandem_repeat_fraction": 0.512,
94
+ "min_tandem_repeat_fraction": 0.0
95
+ },
96
+ "fimo_leonine_js": {
97
+ "scanner": "fimo",
98
+ "motif_universe_size": 2228,
99
+ "pvalue": 0.0001,
100
+ "heatmap": [
101
+ [
102
+ 0.3466716603527195
103
+ ]
104
+ ],
105
+ "cell_types": [
106
+ "Ex"
107
+ ],
108
+ "summary": {
109
+ "diag_mean": 0.3466716603527195,
110
+ "offdiag_mean": 0.0,
111
+ "specificity_ratio": 0.0,
112
+ "n_cell_types": 1
113
+ },
114
+ "frobenius_diff_vs_obs_self": Infinity
115
+ },
116
+ "fimo_ctrldna_motif_corr": {
117
+ "scanner": "fimo",
118
+ "per_cell": {
119
+ "Ex": {
120
+ "pearson_r": 0.41714006519587865,
121
+ "r_squared": 0.1740058339916219,
122
+ "slope": 0.08431869045042495,
123
+ "intercept": 0.000410988020444154,
124
+ "n_motifs": 2228
125
+ }
126
+ },
127
+ "summary": {
128
+ "mean_r_squared": 0.1740058339916219,
129
+ "min_r_squared": 0.1740058339916219,
130
+ "max_r_squared": 0.1740058339916219,
131
+ "n_cells": 1
132
+ }
133
+ },
134
+ "fimo_tf_program_specificity": {
135
+ "scanner": "fimo",
136
+ "per_cell_program_fraction": {
137
+ "Ex": 0.000185359868223917
138
+ },
139
+ "mean_program_fraction": 0.000185359868223917,
140
+ "random_baseline": 0.143,
141
+ "interpretation": "fraction of motif hits in each cell-type's generated enhancers that come from THAT cell type's known TF program (NEUROD/MEF2 for Ex, SPI1/IRF for Mic, etc.). 1/7 = random; higher = better cell-specificity."
142
+ },
143
+ "moods_leonine_js": {
144
+ "scanner": "moods",
145
+ "motif_universe_size": 720,
146
+ "pvalue": 0.0001,
147
+ "heatmap": [
148
+ [
149
+ 0.3472820163811928
150
+ ]
151
+ ],
152
+ "cell_types": [
153
+ "Ex"
154
+ ],
155
+ "summary": {
156
+ "diag_mean": 0.3472820163811928,
157
+ "offdiag_mean": 0.0,
158
+ "specificity_ratio": 0.0,
159
+ "n_cell_types": 1
160
+ },
161
+ "frobenius_diff_vs_obs_self": Infinity
162
+ },
163
+ "moods_ctrldna_motif_corr": {
164
+ "scanner": "moods",
165
+ "per_cell": {
166
+ "Ex": {
167
+ "pearson_r": 0.5058152966811763,
168
+ "r_squared": 0.2558491143566664,
169
+ "slope": 0.08779680255614522,
170
+ "intercept": 0.0012669488853386875,
171
+ "n_motifs": 720
172
+ }
173
+ },
174
+ "summary": {
175
+ "mean_r_squared": 0.2558491143566664,
176
+ "min_r_squared": 0.2558491143566664,
177
+ "max_r_squared": 0.2558491143566664,
178
+ "n_cells": 1
179
+ }
180
+ },
181
+ "moods_tf_program_specificity": {
182
+ "scanner": "moods",
183
+ "per_cell_program_fraction": {
184
+ "Ex": 0.0012070370934367637
185
+ },
186
+ "mean_program_fraction": 0.0012070370934367637,
187
+ "random_baseline": 0.143,
188
+ "interpretation": "fraction of motif hits in each cell-type's generated enhancers that come from THAT cell type's known TF program (NEUROD/MEF2 for Ex, SPI1/IRF for Mic, etc.). 1/7 = random; higher = better cell-specificity."
189
+ },
190
+ "oracle": {
191
+ "oracle_type": "SeparateBinaryOracle",
192
+ "n_scored": 7000,
193
+ "n_aligned_for_fid": 7000,
194
+ "gen_argmax_accuracy": 0.14985714285714286,
195
+ "gen_on_target": -16.296070281074517,
196
+ "gen_off_target": -16.33853605382783,
197
+ "gen_specificity": 0.04246577275331376,
198
+ "gold_argmax_accuracy": 0.9758571428571429,
199
+ "gold_specificity": 26.760889453385555,
200
+ "n_gold": 7000.0,
201
+ "fid": 75.58382929553815,
202
+ "fid_per_cell": {
203
+ "Ex": 119.05023259080659,
204
+ "In": 181.6494814814658,
205
+ "OPC": 102.79811345713617,
206
+ "Ast": 69.39535586995893,
207
+ "Oli": 99.60272615852712,
208
+ "Mic": 73.27570033722156,
209
+ "End": 250.00938061199676
210
+ }
211
+ }
212
+ }
predictions_preview.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
training_log.jsonl ADDED
@@ -0,0 +1,422 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"step": 1, "loss": 0.33768609166145325, "elapsed_s": 0.8, "head": "ar"}
2
+ {"step": 10, "loss": 0.0020572838839143515, "elapsed_s": 2.1, "head": "ar"}
3
+ {"step": 20, "loss": 0.0006668390124104917, "elapsed_s": 3.4, "head": "ar"}
4
+ {"step": 30, "loss": 0.00020201990264467895, "elapsed_s": 4.3, "head": "ar"}
5
+ {"step": 40, "loss": 0.0001245062449015677, "elapsed_s": 5.1, "head": "ar"}
6
+ {"step": 50, "loss": 0.00010640319669619203, "elapsed_s": 6.0, "head": "ar"}
7
+ {"step": 60, "loss": 0.00011547593749128282, "elapsed_s": 6.7, "head": "ar"}
8
+ {"step": 70, "loss": 0.0001431891869287938, "elapsed_s": 8.0, "head": "ar"}
9
+ {"step": 80, "loss": 6.48792993160896e-05, "elapsed_s": 8.9, "head": "ar"}
10
+ {"step": 90, "loss": 0.00044112870818935335, "elapsed_s": 9.8, "head": "ar"}
11
+ {"step": 100, "loss": 5.081587369204499e-05, "elapsed_s": 10.6, "head": "ar"}
12
+ {"step": 110, "loss": 6.372868665494025e-05, "elapsed_s": 11.5, "head": "ar"}
13
+ {"step": 120, "loss": 5.584162499872036e-05, "elapsed_s": 12.8, "head": "ar"}
14
+ {"step": 130, "loss": 0.00022516344324685633, "elapsed_s": 13.8, "head": "ar"}
15
+ {"step": 1, "loss": 0.8157168030738831, "elapsed_s": 1.1, "head": "ar"}
16
+ {"step": 140, "loss": 7.645715231774375e-05, "elapsed_s": 14.6, "head": "ar"}
17
+ {"step": 150, "loss": 0.00011493053170852363, "elapsed_s": 16.9, "head": "ar"}
18
+ {"step": 160, "loss": 7.06012942828238e-05, "elapsed_s": 18.1, "head": "ar"}
19
+ {"step": 10, "loss": 0.5213097333908081, "elapsed_s": 5.3, "head": "ar"}
20
+ {"step": 170, "loss": 0.0005726458039134741, "elapsed_s": 19.3, "head": "ar"}
21
+ {"step": 180, "loss": 0.00010466248932061717, "elapsed_s": 20.7, "head": "ar"}
22
+ {"step": 20, "loss": 0.3670072853565216, "elapsed_s": 7.9, "head": "ar"}
23
+ {"step": 190, "loss": 4.6557550376746804e-05, "elapsed_s": 21.5, "head": "ar"}
24
+ {"step": 200, "loss": 2.7530530132935382e-05, "elapsed_s": 22.6, "head": "ar"}
25
+ {"step": 30, "loss": 0.23992875218391418, "elapsed_s": 10.5, "head": "ar"}
26
+ {"step": 40, "loss": 0.1522972583770752, "elapsed_s": 14.4, "head": "ar"}
27
+ {"step": 50, "loss": 0.10044868290424347, "elapsed_s": 17.6, "head": "ar"}
28
+ {"step": 60, "loss": 0.07074403017759323, "elapsed_s": 20.8, "head": "ar"}
29
+ {"step": 70, "loss": 0.05996284261345863, "elapsed_s": 24.0, "head": "ar"}
30
+ {"kind": "eval", "step": 200, "val_loss": 0.0003785018100918803, "val_n": 7000.0}
31
+ {"step": 210, "loss": 2.609405964903999e-05, "elapsed_s": 40.0, "head": "ar"}
32
+ {"step": 80, "loss": 0.037999916821718216, "elapsed_s": 27.8, "head": "ar"}
33
+ {"step": 220, "loss": 2.1589645257336088e-05, "elapsed_s": 41.6, "head": "ar"}
34
+ {"step": 230, "loss": 2.3236480046762154e-05, "elapsed_s": 42.6, "head": "ar"}
35
+ {"step": 240, "loss": 5.014649286749773e-05, "elapsed_s": 43.5, "head": "ar"}
36
+ {"step": 90, "loss": 0.03023257665336132, "elapsed_s": 31.2, "head": "ar"}
37
+ {"step": 250, "loss": 0.00031191599555313587, "elapsed_s": 44.6, "head": "ar"}
38
+ {"step": 260, "loss": 0.00010953240416711196, "elapsed_s": 45.6, "head": "ar"}
39
+ {"step": 270, "loss": 5.2394778322195634e-05, "elapsed_s": 47.0, "head": "ar"}
40
+ {"step": 100, "loss": 0.025980744510889053, "elapsed_s": 34.5, "head": "ar"}
41
+ {"step": 280, "loss": 0.00010049960837932304, "elapsed_s": 48.5, "head": "ar"}
42
+ {"step": 290, "loss": 7.498916966142133e-05, "elapsed_s": 49.4, "head": "ar"}
43
+ {"step": 110, "loss": 0.0195320937782526, "elapsed_s": 37.1, "head": "ar"}
44
+ {"step": 300, "loss": 2.162451528420206e-05, "elapsed_s": 50.4, "head": "ar"}
45
+ {"step": 310, "loss": 0.00016037160821724683, "elapsed_s": 51.5, "head": "ar"}
46
+ {"step": 320, "loss": 2.1901239961152896e-05, "elapsed_s": 52.5, "head": "ar"}
47
+ {"step": 120, "loss": 0.02047000080347061, "elapsed_s": 40.2, "head": "ar"}
48
+ {"step": 330, "loss": 1.6618041627225466e-05, "elapsed_s": 53.8, "head": "ar"}
49
+ {"step": 340, "loss": 4.1000745113706216e-05, "elapsed_s": 55.0, "head": "ar"}
50
+ {"step": 130, "loss": 0.018111703917384148, "elapsed_s": 42.8, "head": "ar"}
51
+ {"step": 350, "loss": 9.395134838996455e-06, "elapsed_s": 56.2, "head": "ar"}
52
+ {"step": 360, "loss": 1.0806441423483193e-05, "elapsed_s": 57.6, "head": "ar"}
53
+ {"step": 140, "loss": 0.01718146912753582, "elapsed_s": 45.9, "head": "ar"}
54
+ {"step": 370, "loss": 1.0024626135418657e-05, "elapsed_s": 59.1, "head": "ar"}
55
+ {"step": 380, "loss": 8.26425548439147e-06, "elapsed_s": 61.9, "head": "ar"}
56
+ {"step": 150, "loss": 0.013054023496806622, "elapsed_s": 49.3, "head": "ar"}
57
+ {"step": 390, "loss": 1.4594934327760711e-05, "elapsed_s": 62.7, "head": "ar"}
58
+ {"step": 400, "loss": 0.00017360378114972264, "elapsed_s": 63.7, "head": "ar"}
59
+ {"step": 160, "loss": 0.011797660030424595, "elapsed_s": 52.3, "head": "ar"}
60
+ {"step": 170, "loss": 0.010740426369011402, "elapsed_s": 54.8, "head": "ar"}
61
+ {"step": 180, "loss": 0.010314024053514004, "elapsed_s": 57.4, "head": "ar"}
62
+ {"step": 190, "loss": 0.010034145787358284, "elapsed_s": 60.7, "head": "ar"}
63
+ {"step": 200, "loss": 0.00750979408621788, "elapsed_s": 63.6, "head": "ar"}
64
+ {"kind": "eval", "step": 400, "val_loss": 0.00011089673105535829, "val_n": 7000.0}
65
+ {"step": 410, "loss": 0.0016622159164398909, "elapsed_s": 81.9, "head": "ar"}
66
+ {"step": 420, "loss": 6.18189515080303e-05, "elapsed_s": 82.7, "head": "ar"}
67
+ {"step": 430, "loss": 2.8654856578214094e-05, "elapsed_s": 83.9, "head": "ar"}
68
+ {"step": 440, "loss": 2.5333398298243992e-05, "elapsed_s": 84.7, "head": "ar"}
69
+ {"step": 450, "loss": 1.0323085916752461e-05, "elapsed_s": 85.6, "head": "ar"}
70
+ {"step": 460, "loss": 9.699740076030139e-06, "elapsed_s": 87.0, "head": "ar"}
71
+ {"step": 470, "loss": 8.73956651048502e-06, "elapsed_s": 88.5, "head": "ar"}
72
+ {"step": 480, "loss": 7.117927452782169e-06, "elapsed_s": 89.6, "head": "ar"}
73
+ {"step": 490, "loss": 7.005469342402648e-06, "elapsed_s": 91.2, "head": "ar"}
74
+ {"step": 500, "loss": 5.984074050502386e-06, "elapsed_s": 92.1, "head": "ar"}
75
+ {"step": 510, "loss": 3.1450206734007224e-05, "elapsed_s": 93.5, "head": "ar"}
76
+ {"step": 520, "loss": 6.654408934991807e-06, "elapsed_s": 94.8, "head": "ar"}
77
+ {"step": 530, "loss": 6.252863386180252e-06, "elapsed_s": 95.9, "head": "ar"}
78
+ {"step": 540, "loss": 5.476572368934285e-06, "elapsed_s": 97.2, "head": "ar"}
79
+ {"step": 550, "loss": 3.660761649371125e-05, "elapsed_s": 98.5, "head": "ar"}
80
+ {"step": 560, "loss": 5.18998785992153e-06, "elapsed_s": 100.0, "head": "ar"}
81
+ {"step": 570, "loss": 5.317484010447515e-06, "elapsed_s": 101.3, "head": "ar"}
82
+ {"step": 580, "loss": 4.447944593266584e-05, "elapsed_s": 102.6, "head": "ar"}
83
+ {"step": 590, "loss": 3.298748197266832e-05, "elapsed_s": 103.9, "head": "ar"}
84
+ {"step": 600, "loss": 4.920081210002536e-06, "elapsed_s": 105.4, "head": "ar"}
85
+ {"kind": "eval", "step": 600, "val_loss": 1.2555118106839863e-05, "val_n": 7000.0}
86
+ {"step": 610, "loss": 5.3294224926503375e-05, "elapsed_s": 124.8, "head": "ar"}
87
+ {"step": 620, "loss": 5.995512765366584e-06, "elapsed_s": 126.4, "head": "ar"}
88
+ {"step": 630, "loss": 5.4749460105085745e-05, "elapsed_s": 127.4, "head": "ar"}
89
+ {"step": 640, "loss": 4.030084255646216e-06, "elapsed_s": 128.9, "head": "ar"}
90
+ {"step": 650, "loss": 4.0271775105793495e-06, "elapsed_s": 130.5, "head": "ar"}
91
+ {"step": 660, "loss": 4.878559411736205e-05, "elapsed_s": 131.3, "head": "ar"}
92
+ {"step": 670, "loss": 4.225974862492876e-06, "elapsed_s": 132.1, "head": "ar"}
93
+ {"step": 680, "loss": 4.023867859359598e-06, "elapsed_s": 133.5, "head": "ar"}
94
+ {"step": 690, "loss": 2.9697041554754833e-06, "elapsed_s": 134.4, "head": "ar"}
95
+ {"step": 700, "loss": 3.4025149489025353e-06, "elapsed_s": 135.3, "head": "ar"}
96
+ {"step": 710, "loss": 3.6102192098041996e-06, "elapsed_s": 136.2, "head": "ar"}
97
+ {"step": 720, "loss": 1.8541142708272673e-05, "elapsed_s": 137.1, "head": "ar"}
98
+ {"step": 730, "loss": 3.743137858691625e-05, "elapsed_s": 138.6, "head": "ar"}
99
+ {"step": 740, "loss": 3.71455644199159e-06, "elapsed_s": 140.2, "head": "ar"}
100
+ {"step": 750, "loss": 4.635853838408366e-05, "elapsed_s": 141.7, "head": "ar"}
101
+ {"step": 760, "loss": 3.943554929719539e-06, "elapsed_s": 144.8, "head": "ar"}
102
+ {"step": 770, "loss": 8.958242688095197e-06, "elapsed_s": 146.1, "head": "ar"}
103
+ {"step": 780, "loss": 4.242994691594504e-06, "elapsed_s": 147.4, "head": "ar"}
104
+ {"step": 790, "loss": 2.6759439606394153e-06, "elapsed_s": 148.5, "head": "ar"}
105
+ {"step": 800, "loss": 7.50875960875419e-06, "elapsed_s": 149.3, "head": "ar"}
106
+ {"kind": "eval", "step": 800, "val_loss": 9.629162928311935e-06, "val_n": 7000.0}
107
+ {"step": 810, "loss": 2.1677340555470437e-05, "elapsed_s": 169.1, "head": "ar"}
108
+ {"step": 820, "loss": 3.0497349143843167e-06, "elapsed_s": 169.9, "head": "ar"}
109
+ {"step": 830, "loss": 7.272129550983664e-06, "elapsed_s": 170.8, "head": "ar"}
110
+ {"step": 840, "loss": 3.2099267173180124e-06, "elapsed_s": 171.6, "head": "ar"}
111
+ {"step": 850, "loss": 1.3579445294453762e-05, "elapsed_s": 172.9, "head": "ar"}
112
+ {"step": 860, "loss": 2.0844638584094355e-06, "elapsed_s": 173.8, "head": "ar"}
113
+ {"step": 870, "loss": 5.278156095300801e-06, "elapsed_s": 174.6, "head": "ar"}
114
+ {"step": 880, "loss": 1.7874168406706303e-05, "elapsed_s": 175.8, "head": "ar"}
115
+ {"step": 890, "loss": 2.602038875920698e-06, "elapsed_s": 177.1, "head": "ar"}
116
+ {"step": 900, "loss": 2.3132261048885994e-06, "elapsed_s": 179.6, "head": "ar"}
117
+ {"step": 910, "loss": 2.636134695421788e-06, "elapsed_s": 181.2, "head": "ar"}
118
+ {"step": 920, "loss": 3.3073533813876566e-06, "elapsed_s": 182.2, "head": "ar"}
119
+ {"step": 930, "loss": 2.7931755539611913e-06, "elapsed_s": 183.7, "head": "ar"}
120
+ {"step": 940, "loss": 2.4928101538534975e-06, "elapsed_s": 185.1, "head": "ar"}
121
+ {"step": 950, "loss": 4.5148870412958786e-05, "elapsed_s": 186.6, "head": "ar"}
122
+ {"step": 960, "loss": 6.3043489717529155e-06, "elapsed_s": 188.6, "head": "ar"}
123
+ {"step": 970, "loss": 1.2873588275397196e-05, "elapsed_s": 190.8, "head": "ar"}
124
+ {"step": 980, "loss": 8.160362085618544e-06, "elapsed_s": 192.3, "head": "ar"}
125
+ {"step": 990, "loss": 1.246961528522661e-05, "elapsed_s": 193.6, "head": "ar"}
126
+ {"step": 1000, "loss": 3.137994099233765e-06, "elapsed_s": 194.9, "head": "ar"}
127
+ {"kind": "eval", "step": 1000, "val_loss": 1.2657416731599369e-05, "val_n": 7000.0}
128
+ {"step": 1010, "loss": 3.98448673877283e-06, "elapsed_s": 215.9, "head": "ar"}
129
+ {"step": 1020, "loss": 1.0483983714948408e-05, "elapsed_s": 217.3, "head": "ar"}
130
+ {"step": 1030, "loss": 3.134960707029677e-06, "elapsed_s": 219.1, "head": "ar"}
131
+ {"step": 1040, "loss": 7.430694950016914e-06, "elapsed_s": 220.2, "head": "ar"}
132
+ {"step": 1050, "loss": 2.2764318146073492e-06, "elapsed_s": 221.0, "head": "ar"}
133
+ {"step": 1060, "loss": 4.090666152478661e-06, "elapsed_s": 222.5, "head": "ar"}
134
+ {"step": 1070, "loss": 3.77048263544566e-06, "elapsed_s": 223.7, "head": "ar"}
135
+ {"step": 1080, "loss": 2.625357410579454e-06, "elapsed_s": 224.7, "head": "ar"}
136
+ {"step": 1090, "loss": 2.6926776627078652e-05, "elapsed_s": 226.6, "head": "ar"}
137
+ {"step": 1100, "loss": 3.4104432415915653e-06, "elapsed_s": 227.5, "head": "ar"}
138
+ {"step": 1110, "loss": 5.59585532755591e-06, "elapsed_s": 228.3, "head": "ar"}
139
+ {"step": 1120, "loss": 6.753502475476125e-06, "elapsed_s": 229.1, "head": "ar"}
140
+ {"step": 1130, "loss": 1.8659798115550075e-06, "elapsed_s": 230.1, "head": "ar"}
141
+ {"step": 1140, "loss": 0.00010135914635611698, "elapsed_s": 231.0, "head": "ar"}
142
+ {"step": 1150, "loss": 1.5179120964603499e-06, "elapsed_s": 231.8, "head": "ar"}
143
+ {"step": 1160, "loss": 3.1085382943274453e-06, "elapsed_s": 232.9, "head": "ar"}
144
+ {"step": 1170, "loss": 2.2879182779433904e-06, "elapsed_s": 234.1, "head": "ar"}
145
+ {"step": 1180, "loss": 1.4918831539034727e-06, "elapsed_s": 235.0, "head": "ar"}
146
+ {"step": 1190, "loss": 2.450372221574071e-06, "elapsed_s": 236.4, "head": "ar"}
147
+ {"step": 1200, "loss": 1.3282655345392413e-05, "elapsed_s": 237.9, "head": "ar"}
148
+ {"kind": "eval", "step": 200, "val_loss": 0.009116475048874105, "val_n": 7000.0}
149
+ {"step": 210, "loss": 0.00691491924226284, "elapsed_s": 232.0, "head": "ar"}
150
+ {"step": 220, "loss": 0.006404893007129431, "elapsed_s": 235.0, "head": "ar"}
151
+ {"kind": "eval", "step": 1200, "val_loss": 6.336929190963149e-06, "val_n": 7000.0}
152
+ {"step": 1210, "loss": 3.0424314900301397e-05, "elapsed_s": 249.8, "head": "ar"}
153
+ {"step": 1220, "loss": 1.7614775060792454e-06, "elapsed_s": 251.4, "head": "ar"}
154
+ {"step": 230, "loss": 0.006793435197323561, "elapsed_s": 238.5, "head": "ar"}
155
+ {"step": 1230, "loss": 1.8918194655270781e-06, "elapsed_s": 252.3, "head": "ar"}
156
+ {"step": 1240, "loss": 2.116679979735636e-06, "elapsed_s": 253.1, "head": "ar"}
157
+ {"step": 1250, "loss": 2.0416439383552643e-06, "elapsed_s": 254.6, "head": "ar"}
158
+ {"step": 240, "loss": 0.006639133207499981, "elapsed_s": 241.6, "head": "ar"}
159
+ {"step": 1260, "loss": 1.7368615772284102e-06, "elapsed_s": 256.0, "head": "ar"}
160
+ {"step": 1270, "loss": 1.8190287391917082e-06, "elapsed_s": 257.9, "head": "ar"}
161
+ {"step": 250, "loss": 0.005504909437149763, "elapsed_s": 245.2, "head": "ar"}
162
+ {"step": 1280, "loss": 2.0237607714079786e-06, "elapsed_s": 258.9, "head": "ar"}
163
+ {"step": 1290, "loss": 1.84285295290465e-06, "elapsed_s": 259.9, "head": "ar"}
164
+ {"step": 1300, "loss": 2.3768063329043798e-05, "elapsed_s": 260.9, "head": "ar"}
165
+ {"step": 1310, "loss": 1.5927635104162619e-06, "elapsed_s": 261.8, "head": "ar"}
166
+ {"step": 260, "loss": 0.00539375189691782, "elapsed_s": 248.7, "head": "ar"}
167
+ {"step": 1320, "loss": 1.5017211580925505e-06, "elapsed_s": 263.2, "head": "ar"}
168
+ {"step": 1330, "loss": 1.6090239114419091e-06, "elapsed_s": 264.4, "head": "ar"}
169
+ {"step": 270, "loss": 0.005511795170605183, "elapsed_s": 252.2, "head": "ar"}
170
+ {"step": 1340, "loss": 4.480204006540589e-05, "elapsed_s": 265.8, "head": "ar"}
171
+ {"step": 1350, "loss": 3.464040219114395e-06, "elapsed_s": 266.9, "head": "ar"}
172
+ {"step": 1360, "loss": 1.4143442967906594e-06, "elapsed_s": 267.8, "head": "ar"}
173
+ {"step": 280, "loss": 0.007281768601387739, "elapsed_s": 255.3, "head": "ar"}
174
+ {"step": 1370, "loss": 1.3404969649855047e-06, "elapsed_s": 269.1, "head": "ar"}
175
+ {"step": 1380, "loss": 6.660944563918747e-06, "elapsed_s": 270.4, "head": "ar"}
176
+ {"step": 290, "loss": 0.005828922148793936, "elapsed_s": 258.1, "head": "ar"}
177
+ {"step": 1390, "loss": 2.956419166366686e-06, "elapsed_s": 271.7, "head": "ar"}
178
+ {"step": 1400, "loss": 1.8507663526179385e-06, "elapsed_s": 273.4, "head": "ar"}
179
+ {"step": 300, "loss": 0.004460627678781748, "elapsed_s": 261.3, "head": "ar"}
180
+ {"step": 310, "loss": 0.006005497183650732, "elapsed_s": 264.2, "head": "ar"}
181
+ {"step": 320, "loss": 0.004157997202128172, "elapsed_s": 267.6, "head": "ar"}
182
+ {"step": 330, "loss": 0.004970646928995848, "elapsed_s": 271.0, "head": "ar"}
183
+ {"step": 340, "loss": 0.004719988908618689, "elapsed_s": 274.3, "head": "ar"}
184
+ {"step": 350, "loss": 0.0041032820008695126, "elapsed_s": 277.2, "head": "ar"}
185
+ {"kind": "eval", "step": 1400, "val_loss": 5.389277544769097e-06, "val_n": 7000.0}
186
+ {"step": 1410, "loss": 5.817847522848751e-06, "elapsed_s": 292.6, "head": "ar"}
187
+ {"step": 360, "loss": 0.003962608519941568, "elapsed_s": 280.2, "head": "ar"}
188
+ {"step": 1420, "loss": 1.8665193692868343e-06, "elapsed_s": 293.4, "head": "ar"}
189
+ {"step": 1430, "loss": 1.07746780031448e-06, "elapsed_s": 295.0, "head": "ar"}
190
+ {"step": 1440, "loss": 1.3804960872221272e-05, "elapsed_s": 296.1, "head": "ar"}
191
+ {"step": 370, "loss": 0.00427195755764842, "elapsed_s": 283.8, "head": "ar"}
192
+ {"step": 1450, "loss": 2.5708857265271945e-06, "elapsed_s": 297.3, "head": "ar"}
193
+ {"step": 1460, "loss": 1.678746912148199e-06, "elapsed_s": 298.1, "head": "ar"}
194
+ {"step": 1470, "loss": 1.1645178119579214e-06, "elapsed_s": 299.5, "head": "ar"}
195
+ {"step": 380, "loss": 0.003468307200819254, "elapsed_s": 287.0, "head": "ar"}
196
+ {"step": 1480, "loss": 1.3413513215709827e-06, "elapsed_s": 300.9, "head": "ar"}
197
+ {"step": 1490, "loss": 1.460762746319233e-06, "elapsed_s": 302.3, "head": "ar"}
198
+ {"step": 390, "loss": 0.0030817368533462286, "elapsed_s": 289.9, "head": "ar"}
199
+ {"step": 1500, "loss": 3.177382950525498e-06, "elapsed_s": 303.1, "head": "ar"}
200
+ {"step": 1510, "loss": 1.0790473652377841e-06, "elapsed_s": 304.7, "head": "ar"}
201
+ {"step": 1520, "loss": 1.6893155816433136e-06, "elapsed_s": 305.5, "head": "ar"}
202
+ {"step": 400, "loss": 0.002836033934727311, "elapsed_s": 292.9, "head": "ar"}
203
+ {"step": 1530, "loss": 6.527113782794913e-06, "elapsed_s": 307.3, "head": "ar"}
204
+ {"step": 1540, "loss": 3.9027881939546205e-06, "elapsed_s": 308.4, "head": "ar"}
205
+ {"step": 1550, "loss": 1.052610855367675e-06, "elapsed_s": 309.7, "head": "ar"}
206
+ {"step": 1560, "loss": 1.5046135786178638e-06, "elapsed_s": 310.6, "head": "ar"}
207
+ {"step": 1570, "loss": 1.3655255770572694e-06, "elapsed_s": 311.6, "head": "ar"}
208
+ {"step": 1580, "loss": 1.0356235407016356e-06, "elapsed_s": 312.6, "head": "ar"}
209
+ {"step": 1590, "loss": 9.967636742658215e-07, "elapsed_s": 313.5, "head": "ar"}
210
+ {"step": 1600, "loss": 9.797718121262733e-07, "elapsed_s": 314.6, "head": "ar"}
211
+ {"kind": "eval", "step": 1600, "val_loss": 4.5882465413862e-06, "val_n": 7000.0}
212
+ {"step": 1610, "loss": 1.982865342142759e-06, "elapsed_s": 335.1, "head": "ar"}
213
+ {"step": 1620, "loss": 1.0667957894838764e-06, "elapsed_s": 335.9, "head": "ar"}
214
+ {"step": 1630, "loss": 9.836680874286685e-07, "elapsed_s": 337.3, "head": "ar"}
215
+ {"step": 1640, "loss": 1.1697953823386342e-06, "elapsed_s": 338.2, "head": "ar"}
216
+ {"step": 1650, "loss": 9.669865903561004e-07, "elapsed_s": 339.7, "head": "ar"}
217
+ {"step": 1660, "loss": 1.6609818658253062e-06, "elapsed_s": 340.8, "head": "ar"}
218
+ {"step": 1670, "loss": 8.009298653632868e-06, "elapsed_s": 342.6, "head": "ar"}
219
+ {"step": 1680, "loss": 1.0714998097682837e-05, "elapsed_s": 343.4, "head": "ar"}
220
+ {"step": 1690, "loss": 1.0042360827355878e-06, "elapsed_s": 344.3, "head": "ar"}
221
+ {"step": 1700, "loss": 1.7606315623197588e-06, "elapsed_s": 346.0, "head": "ar"}
222
+ {"step": 1710, "loss": 8.456366913378588e-07, "elapsed_s": 347.3, "head": "ar"}
223
+ {"step": 1720, "loss": 1.0064937896459014e-06, "elapsed_s": 348.6, "head": "ar"}
224
+ {"step": 1730, "loss": 2.980214730996522e-06, "elapsed_s": 349.4, "head": "ar"}
225
+ {"step": 1740, "loss": 1.628143945708871e-05, "elapsed_s": 350.4, "head": "ar"}
226
+ {"step": 1750, "loss": 1.38073346533929e-05, "elapsed_s": 352.0, "head": "ar"}
227
+ {"step": 1760, "loss": 2.616614438011311e-06, "elapsed_s": 353.7, "head": "ar"}
228
+ {"step": 1770, "loss": 1.0966933814415825e-06, "elapsed_s": 355.2, "head": "ar"}
229
+ {"step": 1780, "loss": 1.0247426871501375e-05, "elapsed_s": 356.7, "head": "ar"}
230
+ {"step": 1790, "loss": 8.270681632893684e-07, "elapsed_s": 358.2, "head": "ar"}
231
+ {"step": 1800, "loss": 7.706809128649184e-07, "elapsed_s": 359.9, "head": "ar"}
232
+ {"kind": "eval", "step": 1800, "val_loss": 3.971535916012467e-06, "val_n": 7000.0}
233
+ {"step": 1810, "loss": 9.015690238811658e-07, "elapsed_s": 379.3, "head": "ar"}
234
+ {"step": 1820, "loss": 1.1889211464222171e-06, "elapsed_s": 380.0, "head": "ar"}
235
+ {"step": 1830, "loss": 1.0762398687802488e-06, "elapsed_s": 380.9, "head": "ar"}
236
+ {"step": 1840, "loss": 1.1680751413223334e-05, "elapsed_s": 382.3, "head": "ar"}
237
+ {"step": 1850, "loss": 8.738551855458354e-07, "elapsed_s": 384.1, "head": "ar"}
238
+ {"step": 1860, "loss": 7.443987897204352e-07, "elapsed_s": 385.2, "head": "ar"}
239
+ {"step": 1870, "loss": 1.306151943936129e-06, "elapsed_s": 386.1, "head": "ar"}
240
+ {"step": 1880, "loss": 9.60975739872083e-06, "elapsed_s": 387.3, "head": "ar"}
241
+ {"step": 1890, "loss": 2.0746436348417774e-05, "elapsed_s": 389.5, "head": "ar"}
242
+ {"step": 1900, "loss": 8.607777090219315e-06, "elapsed_s": 390.5, "head": "ar"}
243
+ {"step": 1910, "loss": 8.609961810179811e-07, "elapsed_s": 391.4, "head": "ar"}
244
+ {"step": 1920, "loss": 1.301057409364148e-06, "elapsed_s": 392.9, "head": "ar"}
245
+ {"step": 1930, "loss": 6.87683268552064e-07, "elapsed_s": 394.1, "head": "ar"}
246
+ {"step": 1940, "loss": 9.049733193933207e-07, "elapsed_s": 395.2, "head": "ar"}
247
+ {"step": 1950, "loss": 6.633378575315874e-07, "elapsed_s": 397.1, "head": "ar"}
248
+ {"step": 1960, "loss": 1.714162181087886e-06, "elapsed_s": 398.3, "head": "ar"}
249
+ {"step": 1970, "loss": 4.80768721899949e-06, "elapsed_s": 400.0, "head": "ar"}
250
+ {"step": 1980, "loss": 4.8893061830312945e-06, "elapsed_s": 401.2, "head": "ar"}
251
+ {"step": 1990, "loss": 7.843904654691869e-07, "elapsed_s": 402.6, "head": "ar"}
252
+ {"step": 2000, "loss": 6.791338819311932e-06, "elapsed_s": 404.2, "head": "ar"}
253
+ {"kind": "eval", "step": 2000, "val_loss": 3.5206360957350365e-06, "val_n": 7000.0}
254
+ {"kind": "eval", "step": 400, "val_loss": 0.0036485758289428695, "val_n": 7000.0}
255
+ {"step": 410, "loss": 0.0038183745928108692, "elapsed_s": 439.5, "head": "ar"}
256
+ {"step": 420, "loss": 0.0026294304989278316, "elapsed_s": 441.8, "head": "ar"}
257
+ {"step": 430, "loss": 0.003248143708333373, "elapsed_s": 444.4, "head": "ar"}
258
+ {"step": 440, "loss": 0.0027418595273047686, "elapsed_s": 447.2, "head": "ar"}
259
+ {"step": 450, "loss": 0.0025944842491298914, "elapsed_s": 450.0, "head": "ar"}
260
+ {"step": 460, "loss": 0.002990575972944498, "elapsed_s": 452.7, "head": "ar"}
261
+ {"step": 470, "loss": 0.0027225736994296312, "elapsed_s": 455.3, "head": "ar"}
262
+ {"step": 480, "loss": 0.002925804117694497, "elapsed_s": 458.1, "head": "ar"}
263
+ {"step": 490, "loss": 0.0024290902074426413, "elapsed_s": 460.6, "head": "ar"}
264
+ {"step": 500, "loss": 0.002367342822253704, "elapsed_s": 463.7, "head": "ar"}
265
+ {"step": 510, "loss": 0.003480238141492009, "elapsed_s": 466.2, "head": "ar"}
266
+ {"step": 520, "loss": 0.0022945234086364508, "elapsed_s": 469.3, "head": "ar"}
267
+ {"step": 530, "loss": 0.0027580824680626392, "elapsed_s": 472.1, "head": "ar"}
268
+ {"step": 540, "loss": 0.0018581290496513247, "elapsed_s": 474.3, "head": "ar"}
269
+ {"step": 550, "loss": 0.0022819642908871174, "elapsed_s": 476.7, "head": "ar"}
270
+ {"step": 560, "loss": 0.002386125037446618, "elapsed_s": 479.4, "head": "ar"}
271
+ {"step": 570, "loss": 0.0019037285819649696, "elapsed_s": 482.1, "head": "ar"}
272
+ {"step": 580, "loss": 0.002807535231113434, "elapsed_s": 484.5, "head": "ar"}
273
+ {"step": 590, "loss": 0.003186720423400402, "elapsed_s": 486.8, "head": "ar"}
274
+ {"step": 600, "loss": 0.0017188055207952857, "elapsed_s": 489.3, "head": "ar"}
275
+ {"kind": "eval", "step": 600, "val_loss": 0.002241770782108818, "val_n": 7000.0}
276
+ {"step": 610, "loss": 0.003322844160720706, "elapsed_s": 818.3, "head": "ar"}
277
+ {"step": 620, "loss": 0.002120278775691986, "elapsed_s": 820.5, "head": "ar"}
278
+ {"step": 630, "loss": 0.0027013998478651047, "elapsed_s": 823.6, "head": "ar"}
279
+ {"step": 640, "loss": 0.0025664535351097584, "elapsed_s": 827.3, "head": "ar"}
280
+ {"step": 650, "loss": 0.0016682040877640247, "elapsed_s": 830.1, "head": "ar"}
281
+ {"step": 660, "loss": 0.0025910725817084312, "elapsed_s": 832.6, "head": "ar"}
282
+ {"step": 670, "loss": 0.0012992571573704481, "elapsed_s": 835.6, "head": "ar"}
283
+ {"step": 680, "loss": 0.0017818965716287494, "elapsed_s": 838.2, "head": "ar"}
284
+ {"step": 690, "loss": 0.0015026392647996545, "elapsed_s": 840.9, "head": "ar"}
285
+ {"step": 700, "loss": 0.0015572598204016685, "elapsed_s": 843.9, "head": "ar"}
286
+ {"step": 710, "loss": 0.001492018229328096, "elapsed_s": 846.8, "head": "ar"}
287
+ {"step": 720, "loss": 0.002090647118166089, "elapsed_s": 849.9, "head": "ar"}
288
+ {"step": 730, "loss": 0.0016441759653389454, "elapsed_s": 853.1, "head": "ar"}
289
+ {"step": 740, "loss": 0.0012522173346951604, "elapsed_s": 855.6, "head": "ar"}
290
+ {"step": 750, "loss": 0.003108511446043849, "elapsed_s": 858.5, "head": "ar"}
291
+ {"step": 760, "loss": 0.0014899007510393858, "elapsed_s": 861.4, "head": "ar"}
292
+ {"step": 770, "loss": 0.0017582301516085863, "elapsed_s": 863.6, "head": "ar"}
293
+ {"step": 780, "loss": 0.001970557263121009, "elapsed_s": 866.1, "head": "ar"}
294
+ {"step": 790, "loss": 0.000990280881524086, "elapsed_s": 868.7, "head": "ar"}
295
+ {"step": 800, "loss": 0.0015064387116581202, "elapsed_s": 871.5, "head": "ar"}
296
+ {"kind": "eval", "step": 800, "val_loss": 0.001680548977838563, "val_n": 7000.0}
297
+ {"step": 810, "loss": 0.002819423098117113, "elapsed_s": 1187.4, "head": "ar"}
298
+ {"step": 820, "loss": 0.0011394670000299811, "elapsed_s": 1190.5, "head": "ar"}
299
+ {"step": 830, "loss": 0.0012815529480576515, "elapsed_s": 1193.1, "head": "ar"}
300
+ {"step": 840, "loss": 0.001399407396093011, "elapsed_s": 1195.6, "head": "ar"}
301
+ {"step": 850, "loss": 0.0024891591165214777, "elapsed_s": 1198.5, "head": "ar"}
302
+ {"step": 860, "loss": 0.0011431269813328981, "elapsed_s": 1201.2, "head": "ar"}
303
+ {"step": 870, "loss": 0.0020568177569657564, "elapsed_s": 1203.7, "head": "ar"}
304
+ {"step": 880, "loss": 0.0030989721417427063, "elapsed_s": 1206.4, "head": "ar"}
305
+ {"step": 890, "loss": 0.0018834383226931095, "elapsed_s": 1208.9, "head": "ar"}
306
+ {"step": 900, "loss": 0.0010720593854784966, "elapsed_s": 1211.6, "head": "ar"}
307
+ {"step": 910, "loss": 0.0010234791552647948, "elapsed_s": 1214.4, "head": "ar"}
308
+ {"step": 920, "loss": 0.002013811143115163, "elapsed_s": 1216.9, "head": "ar"}
309
+ {"step": 930, "loss": 0.0011474337661638856, "elapsed_s": 1219.8, "head": "ar"}
310
+ {"step": 940, "loss": 0.0009459101129323244, "elapsed_s": 1222.5, "head": "ar"}
311
+ {"step": 950, "loss": 0.0011496401857584715, "elapsed_s": 1225.6, "head": "ar"}
312
+ {"step": 960, "loss": 0.0012888332130387425, "elapsed_s": 1228.1, "head": "ar"}
313
+ {"step": 970, "loss": 0.002265971153974533, "elapsed_s": 1231.4, "head": "ar"}
314
+ {"step": 980, "loss": 0.0010454916628077626, "elapsed_s": 1233.8, "head": "ar"}
315
+ {"step": 990, "loss": 0.000888589711394161, "elapsed_s": 1236.6, "head": "ar"}
316
+ {"step": 1000, "loss": 0.000993296504020691, "elapsed_s": 1239.4, "head": "ar"}
317
+ {"kind": "eval", "step": 1000, "val_loss": 0.0013805327682223703, "val_n": 7000.0}
318
+ {"step": 1010, "loss": 0.0010315170511603355, "elapsed_s": 1518.0, "head": "ar"}
319
+ {"step": 1020, "loss": 0.001375523628666997, "elapsed_s": 1520.6, "head": "ar"}
320
+ {"step": 1030, "loss": 0.0010681935818865895, "elapsed_s": 1522.9, "head": "ar"}
321
+ {"step": 1040, "loss": 0.0021271747536957264, "elapsed_s": 1526.1, "head": "ar"}
322
+ {"step": 1050, "loss": 0.0007033874862827361, "elapsed_s": 1528.8, "head": "ar"}
323
+ {"step": 1060, "loss": 0.0011003401596099138, "elapsed_s": 1532.1, "head": "ar"}
324
+ {"step": 1070, "loss": 0.0013504636008292437, "elapsed_s": 1534.3, "head": "ar"}
325
+ {"step": 1080, "loss": 0.0009967918740585446, "elapsed_s": 1537.4, "head": "ar"}
326
+ {"step": 1090, "loss": 0.003230726346373558, "elapsed_s": 1540.0, "head": "ar"}
327
+ {"step": 1100, "loss": 0.001059171394445002, "elapsed_s": 1542.7, "head": "ar"}
328
+ {"step": 1110, "loss": 0.0011096205562353134, "elapsed_s": 1545.0, "head": "ar"}
329
+ {"step": 1120, "loss": 0.001244153711013496, "elapsed_s": 1547.6, "head": "ar"}
330
+ {"step": 1130, "loss": 0.0006119530298747122, "elapsed_s": 1550.1, "head": "ar"}
331
+ {"step": 1140, "loss": 0.00550880329683423, "elapsed_s": 1552.6, "head": "ar"}
332
+ {"step": 1150, "loss": 0.0007053437293507159, "elapsed_s": 1554.8, "head": "ar"}
333
+ {"step": 1160, "loss": 0.001157013582997024, "elapsed_s": 1557.6, "head": "ar"}
334
+ {"step": 1170, "loss": 0.0009195170132443309, "elapsed_s": 1559.8, "head": "ar"}
335
+ {"step": 1180, "loss": 0.0008319061598740518, "elapsed_s": 1562.6, "head": "ar"}
336
+ {"step": 1190, "loss": 0.0008269947720691562, "elapsed_s": 1565.1, "head": "ar"}
337
+ {"step": 1200, "loss": 0.0012492763344198465, "elapsed_s": 1567.9, "head": "ar"}
338
+ {"kind": "eval", "step": 1200, "val_loss": 0.0012779013608316225, "val_n": 7000.0}
339
+ {"step": 1210, "loss": 0.0018258931813761592, "elapsed_s": 1874.3, "head": "ar"}
340
+ {"step": 1220, "loss": 0.0009055273258127272, "elapsed_s": 1876.8, "head": "ar"}
341
+ {"step": 1230, "loss": 0.0008226312347687781, "elapsed_s": 1879.5, "head": "ar"}
342
+ {"step": 1240, "loss": 0.0009587894892320037, "elapsed_s": 1882.4, "head": "ar"}
343
+ {"step": 1250, "loss": 0.0011874581687152386, "elapsed_s": 1885.5, "head": "ar"}
344
+ {"step": 1260, "loss": 0.000995468464680016, "elapsed_s": 1888.2, "head": "ar"}
345
+ {"step": 1270, "loss": 0.0011743544600903988, "elapsed_s": 1891.4, "head": "ar"}
346
+ {"step": 1280, "loss": 0.000934170326218009, "elapsed_s": 1894.4, "head": "ar"}
347
+ {"step": 1290, "loss": 0.0013409317471086979, "elapsed_s": 1897.3, "head": "ar"}
348
+ {"step": 1300, "loss": 0.001514748320914805, "elapsed_s": 1899.9, "head": "ar"}
349
+ {"step": 1310, "loss": 0.0010219430550932884, "elapsed_s": 1903.0, "head": "ar"}
350
+ {"step": 1320, "loss": 0.0009012930677272379, "elapsed_s": 1906.0, "head": "ar"}
351
+ {"step": 1330, "loss": 0.0012672283919528127, "elapsed_s": 1909.8, "head": "ar"}
352
+ {"step": 1340, "loss": 0.0013146441197022796, "elapsed_s": 1912.9, "head": "ar"}
353
+ {"step": 1350, "loss": 0.0016613602638244629, "elapsed_s": 1915.4, "head": "ar"}
354
+ {"step": 1360, "loss": 0.000870061048772186, "elapsed_s": 1918.0, "head": "ar"}
355
+ {"step": 1370, "loss": 0.0009106548968702555, "elapsed_s": 1920.5, "head": "ar"}
356
+ {"step": 1380, "loss": 0.0014916631625965238, "elapsed_s": 1923.0, "head": "ar"}
357
+ {"step": 1390, "loss": 0.0011943557765334845, "elapsed_s": 1926.3, "head": "ar"}
358
+ {"step": 1400, "loss": 0.0009589027031324804, "elapsed_s": 1929.4, "head": "ar"}
359
+ {"kind": "eval", "step": 1400, "val_loss": 0.00113910235707382, "val_n": 7000.0}
360
+ {"step": 1410, "loss": 0.002466773148626089, "elapsed_s": 2240.3, "head": "ar"}
361
+ {"step": 1420, "loss": 0.0009186464594677091, "elapsed_s": 2243.6, "head": "ar"}
362
+ {"step": 1430, "loss": 0.0005913503118790686, "elapsed_s": 2247.2, "head": "ar"}
363
+ {"step": 1440, "loss": 0.0012442157603800297, "elapsed_s": 2250.7, "head": "ar"}
364
+ {"step": 1450, "loss": 0.0010779427830129862, "elapsed_s": 2254.4, "head": "ar"}
365
+ {"step": 1460, "loss": 0.0011523909633979201, "elapsed_s": 2257.5, "head": "ar"}
366
+ {"step": 1470, "loss": 0.0007289984496310353, "elapsed_s": 2261.3, "head": "ar"}
367
+ {"step": 1480, "loss": 0.0012722192332148552, "elapsed_s": 2265.2, "head": "ar"}
368
+ {"step": 1490, "loss": 0.0009974830318242311, "elapsed_s": 2268.7, "head": "ar"}
369
+ {"step": 1500, "loss": 0.000930821115616709, "elapsed_s": 2271.5, "head": "ar"}
370
+ {"step": 1510, "loss": 0.0005651878891512752, "elapsed_s": 2275.0, "head": "ar"}
371
+ {"step": 1520, "loss": 0.0006382225546985865, "elapsed_s": 2278.0, "head": "ar"}
372
+ {"step": 1530, "loss": 0.0010055321035906672, "elapsed_s": 2281.3, "head": "ar"}
373
+ {"step": 1540, "loss": 0.0012934308033436537, "elapsed_s": 2284.7, "head": "ar"}
374
+ {"step": 1550, "loss": 0.000664828869048506, "elapsed_s": 2287.9, "head": "ar"}
375
+ {"step": 1560, "loss": 0.0008923231507651508, "elapsed_s": 2290.7, "head": "ar"}
376
+ {"step": 1570, "loss": 0.0006905746995471418, "elapsed_s": 2294.1, "head": "ar"}
377
+ {"step": 1580, "loss": 0.0006678836070932448, "elapsed_s": 2298.0, "head": "ar"}
378
+ {"step": 1590, "loss": 0.0006206026300787926, "elapsed_s": 2301.5, "head": "ar"}
379
+ {"step": 1600, "loss": 0.0006041897577233613, "elapsed_s": 2304.9, "head": "ar"}
380
+ {"kind": "eval", "step": 1600, "val_loss": 0.001002927115187049, "val_n": 7000.0}
381
+ {"step": 1610, "loss": 0.000986228697001934, "elapsed_s": 2628.1, "head": "ar"}
382
+ {"step": 1620, "loss": 0.0007126597338356078, "elapsed_s": 2631.2, "head": "ar"}
383
+ {"step": 1630, "loss": 0.0007028065738268197, "elapsed_s": 2633.7, "head": "ar"}
384
+ {"step": 1640, "loss": 0.0007537198252975941, "elapsed_s": 2636.6, "head": "ar"}
385
+ {"step": 1650, "loss": 0.0005334328161552548, "elapsed_s": 2639.6, "head": "ar"}
386
+ {"step": 1660, "loss": 0.0010687893955036998, "elapsed_s": 2641.8, "head": "ar"}
387
+ {"step": 1670, "loss": 0.0008954349905252457, "elapsed_s": 2644.7, "head": "ar"}
388
+ {"step": 1680, "loss": 0.0009879261488094926, "elapsed_s": 2647.6, "head": "ar"}
389
+ {"step": 1690, "loss": 0.0008526460151188076, "elapsed_s": 2650.1, "head": "ar"}
390
+ {"step": 1700, "loss": 0.0006849347264505923, "elapsed_s": 2653.7, "head": "ar"}
391
+ {"step": 1710, "loss": 0.0006179868942126632, "elapsed_s": 2656.3, "head": "ar"}
392
+ {"step": 1720, "loss": 0.0007618499803356826, "elapsed_s": 2660.2, "head": "ar"}
393
+ {"step": 1730, "loss": 0.0007457428146153688, "elapsed_s": 2663.2, "head": "ar"}
394
+ {"step": 1740, "loss": 0.0014043848495930433, "elapsed_s": 2666.6, "head": "ar"}
395
+ {"step": 1750, "loss": 0.0014561469433829188, "elapsed_s": 2670.0, "head": "ar"}
396
+ {"step": 1760, "loss": 0.0011086587328463793, "elapsed_s": 2673.3, "head": "ar"}
397
+ {"step": 1770, "loss": 0.0006323332199826837, "elapsed_s": 2676.3, "head": "ar"}
398
+ {"step": 1780, "loss": 0.0015908291097730398, "elapsed_s": 2679.9, "head": "ar"}
399
+ {"step": 1790, "loss": 0.0005683208582922816, "elapsed_s": 2683.5, "head": "ar"}
400
+ {"step": 1800, "loss": 0.0004560278612188995, "elapsed_s": 2687.3, "head": "ar"}
401
+ {"kind": "eval", "step": 1800, "val_loss": 0.0010068494210219276, "val_n": 7000.0}
402
+ {"step": 1810, "loss": 0.0007359150913543999, "elapsed_s": 2970.8, "head": "ar"}
403
+ {"step": 1820, "loss": 0.00099342898465693, "elapsed_s": 2973.9, "head": "ar"}
404
+ {"step": 1830, "loss": 0.0007030203123576939, "elapsed_s": 2977.1, "head": "ar"}
405
+ {"step": 1840, "loss": 0.0010482734069228172, "elapsed_s": 2979.7, "head": "ar"}
406
+ {"step": 1850, "loss": 0.0008859018562361598, "elapsed_s": 2982.5, "head": "ar"}
407
+ {"step": 1860, "loss": 0.0005704698851332068, "elapsed_s": 2986.0, "head": "ar"}
408
+ {"step": 1870, "loss": 0.0007724370807409286, "elapsed_s": 2988.6, "head": "ar"}
409
+ {"step": 1880, "loss": 0.0010715826647356153, "elapsed_s": 2991.0, "head": "ar"}
410
+ {"step": 1890, "loss": 0.001849786494858563, "elapsed_s": 2993.8, "head": "ar"}
411
+ {"step": 1900, "loss": 0.000943592400290072, "elapsed_s": 2996.4, "head": "ar"}
412
+ {"step": 1910, "loss": 0.00047185056610032916, "elapsed_s": 2999.1, "head": "ar"}
413
+ {"step": 1920, "loss": 0.00078765390207991, "elapsed_s": 3002.3, "head": "ar"}
414
+ {"step": 1930, "loss": 0.000458514696219936, "elapsed_s": 3004.9, "head": "ar"}
415
+ {"step": 1940, "loss": 0.0005791429430246353, "elapsed_s": 3007.3, "head": "ar"}
416
+ {"step": 1950, "loss": 0.0006057161372154951, "elapsed_s": 3010.1, "head": "ar"}
417
+ {"step": 1960, "loss": 0.0008873195620253682, "elapsed_s": 3012.8, "head": "ar"}
418
+ {"step": 1970, "loss": 0.0020482316613197327, "elapsed_s": 3015.6, "head": "ar"}
419
+ {"step": 1980, "loss": 0.000935696589294821, "elapsed_s": 3019.0, "head": "ar"}
420
+ {"step": 1990, "loss": 0.0007570537854917347, "elapsed_s": 3022.3, "head": "ar"}
421
+ {"step": 2000, "loss": 0.0013412516564130783, "elapsed_s": 3025.4, "head": "ar"}
422
+ {"kind": "eval", "step": 2000, "val_loss": 0.0008683186207365777, "val_n": 7000.0}