File tree Expand file tree Collapse file tree
Expand file tree Collapse file tree Original file line number Diff line number Diff line change 1+ {
2+ "schema_version" : 1 ,
3+ "kind" : " adr_0011_toy_prototype_g_x1" ,
4+ "config" : {
5+ "model" : " google/gemma-3-1b-it" ,
6+ "device" : " mps" ,
7+ "cross_attn_depth" : 8 ,
8+ "sink" : 4 ,
9+ "window" : 64 ,
10+ "num_heads" : 8 ,
11+ "head_dim" : 64 ,
12+ "train_steps" : 200 ,
13+ "lr" : 0.0003 ,
14+ "n_train" : 200 ,
15+ "n_eval" : 50 ,
16+ "haystack_min_tokens" : 256 ,
17+ "haystack_max_tokens" : 1024 ,
18+ "seed" : 42
19+ },
20+ "pre_train" : {
21+ "cross_attn_recall" : 0.0 ,
22+ "baseline_recall" : 0.0
23+ },
24+ "training_history" : [
25+ {
26+ "step" : 50 ,
27+ "cross_attn_recall" : 0.0 ,
28+ "baseline_recall" : 0.0 ,
29+ "loss_avg10" : 65.85
30+ },
31+ {
32+ "step" : 100 ,
33+ "cross_attn_recall" : 0.0 ,
34+ "baseline_recall" : 0.0 ,
35+ "loss_avg10" : 31.575
36+ },
37+ {
38+ "step" : 150 ,
39+ "cross_attn_recall" : 0.0 ,
40+ "baseline_recall" : 0.0 ,
41+ "loss_avg10" : 21.975
42+ },
43+ {
44+ "step" : 200 ,
45+ "cross_attn_recall" : 0.0 ,
46+ "baseline_recall" : 0.0 ,
47+ "loss_avg10" : 23.24375
48+ }
49+ ],
50+ "final" : {
51+ "cross_attn_recall" : 0.0 ,
52+ "baseline_recall" : 0.0 ,
53+ "elapsed_s" : 1698.5560488329502
54+ },
55+ "gate_g_x1_pass" : false
56+ }
You can’t perform that action at this time.
0 commit comments