Upload training_logs/hill_z_ranked_T20_s5__seed42.json with huggingface_hub

training_logs/hill_z_ranked_T20_s5__seed42.json

@@ -0,0 +1,100 @@
+{
+  "loss": [
+    105360.046875,
+    50075.2109375,
+    44736.578125,
+    42076.578125,
+    40220.83203125,
+    38995.0,
+    38201.90234375,
+    37484.3203125,
+    37149.1953125,
+    36867.93359375,
+    36508.76953125,
+    36258.078125,
+    36260.0078125,
+    36153.71484375,
+    35894.4609375,
+    35914.6796875,
+    35824.125,
+    35687.9921875
+  ],
+  "l0": [
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0,
+    500.0
+  ],
+  "steps_logged": [
+    0,
+    200,
+    400,
+    600,
+    800,
+    1000,
+    1200,
+    1400,
+    1600,
+    1800,
+    2000,
+    2200,
+    2400,
+    2600,
+    2800,
+    3000,
+    3200,
+    3400
+  ],
+  "final_step": 3400,
+  "converged": true,
+  "plateau_last": 0.01949810987198692,
+  "elapsed_s": 2601.677392244339,
+  "T_max": 20,
+  "shifts": null,
+  "matryoshka_h_size": 3686,
+  "alpha": 1.0,
+  "src_class": "SubseqRankedH8",
+  "t_sample": 5,
+  "n_seqs_used": 24000,
+  "ctx_used": 64,
+  "ctx_slice_direction": "last",
+  "row": 605,
+  "arch_id": "hill_z_ranked_T20_s5",
+  "arch": "hill_z_ranked_T20_s5",
+  "group": 99,
+  "src_module": "src.architectures.phase7_subseq_z_variants",
+  "T": null,
+  "n_layers": null,
+  "k_win": 500,
+  "k_pos": 25,
+  "gamma": null,
+  "n_scales": null,
+  "seed": 42,
+  "d_in": 2304,
+  "d_sae": 18432,
+  "subject_model": "google/gemma-2-2b",
+  "anchor_layer": 12,
+  "mlc_layers": [
+    10,
+    11,
+    12,
+    13,
+    14
+  ],
+  "phase": "phase7_unification",
+  "run_id": "hill_z_ranked_T20_s5__seed42"
+}